zotero/resource/bluebird.js

5400 lines
167 KiB
JavaScript
Raw Normal View History

/* @preserve
* The MIT License (MIT)
*
2015-10-29 06:23:22 +00:00
* Copyright (c) 2013-2015 Petka Antonov
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
2015-10-29 06:23:22 +00:00
* furnished to do so, subject to the following conditions:
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*
*/
/**
2016-01-09 02:07:10 +00:00
* bluebird build version 3.1.1
* Features enabled: core, race, call_get, generators, map, nodeify, promisify, props, reduce, settle, some, using, timers, filter, any, each
*/
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
!function(e){
2015-10-29 06:23:22 +00:00
//
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
// Added by Zotero
//
EXPORTED_SYMBOLS = ["Promise"];
// Set BackstagePass (which contains .Error, etc.) as global object
self = this;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
// We need to maintain references to running nsITimers. Otherwise, they can
// get garbage collected before they fire. Also, we might need to cancel them.
var _runningTimers = {};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
// Provide an implementation of setTimeout
self.setTimeout = function (func, ms) {
var id = Math.floor(Math.random() * (1000000000000 - 1)) + 1
var useMethodjit = Components.utils.methodjit;
var timer = Components.classes["@mozilla.org/timer;1"]
.createInstance(Components.interfaces.nsITimer);
timer.initWithCallback({"notify":function() {
// Remove timer from object so it can be garbage collected
delete _runningTimers[id];
// Execute callback function
try {
func();
} catch(err) {
// Rethrow errors that occur so that they appear in the error
// console with the appropriate name and line numbers. While the
// the errors appear without this, the line numbers get eaten.
var scriptError = Components.classes["@mozilla.org/scripterror;1"]
.createInstance(Components.interfaces.nsIScriptError);
scriptError.init(
err.message || err.toString(),
err.fileName || err.filename || null,
null,
err.lineNumber || null,
null,
scriptError.errorFlag,
'component javascript'
);
Components.classes["@mozilla.org/consoleservice;1"]
.getService(Components.interfaces.nsIConsoleService)
.logMessage(scriptError);
self.debug(err.stack, 1);
}
}}, ms, Components.interfaces.nsITimer.TYPE_ONE_SHOT);
_runningTimers[id] = timer;
return id;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
self.clearTimeout = function (id) {
var timer = _runningTimers[id];
if (timer) {
timer.cancel();
}
delete _runningTimers[id];
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
self.debug = function (msg) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
dump(msg + "\n\n");
};
Promise = e();
2016-01-09 02:07:10 +00:00
Promise.config({
warnings: true,
longStackTraces: true,
cancellation: true
});
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
// TEMP: Only turn on if debug logging enabled?
Promise.onPossiblyUnhandledRejection(function (e, promise) {
if (e.name == 'ZoteroPromiseInterrupt') {
return;
}
// Ignore some errors during tests
if (e.message && e.message.indexOf(' -- ignore') != -1) return;
2015-10-29 06:23:22 +00:00
self.debug('Possibly unhandled rejection:\n\n' + e.stack);
throw e;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
});
return;
2016-01-09 02:07:10 +00:00
if("object"==typeof exports&&"undefined"!=typeof module)module.exports=e();else if("function"==typeof define&&define.amd)define([],e);else{var f;"undefined"!=typeof window?f=window:"undefined"!=typeof global?f=global:"undefined"!=typeof self&&(f=self),f.Promise=e()}}(function(){var define,module,exports;return (function e(t,n,r){function s(o,u){if(!n[o]){if(!t[o]){var a=typeof _dereq_=="function"&&_dereq_;if(!u&&a)return a(o,!0);if(i)return i(o,!0);var f=new Error("Cannot find module '"+o+"'");throw f.code="MODULE_NOT_FOUND",f}var l=n[o]={exports:{}};t[o][0].call(l.exports,function(e){var n=t[o][1][e];return s(n?n:e)},l,l.exports,e,t,n,r)}return n[o].exports}var i=typeof _dereq_=="function"&&_dereq_;for(var o=0;o<r.length;o++)s(r[o]);return s})({1:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise) {
var SomePromiseArray = Promise._SomePromiseArray;
function any(promises) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = new SomePromiseArray(promises);
var promise = ret.promise();
ret.setHowMany(1);
ret.setUnwrap();
ret.init();
return promise;
}
Promise.any = function (promises) {
return any(promises);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.any = function () {
return any(this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
},{}],2:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
var firstLineError;
try {throw new Error(); } catch (e) {firstLineError = e;}
2016-01-09 02:07:10 +00:00
var schedule = _dereq_("./schedule");
var Queue = _dereq_("./queue");
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function Async() {
this._isTickUsed = false;
this._lateQueue = new Queue(16);
this._normalQueue = new Queue(16);
2016-01-09 02:07:10 +00:00
this._haveDrainedQueues = false;
2015-06-08 06:38:56 +00:00
this._trampolineEnabled = true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var self = this;
this.drainQueues = function () {
self._drainQueues();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
this._schedule = schedule;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
Async.prototype.enableTrampoline = function() {
this._trampolineEnabled = true;
};
2015-06-08 06:38:56 +00:00
Async.prototype.disableTrampolineIfNecessary = function() {
if (util.hasDevTools) {
this._trampolineEnabled = false;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Async.prototype.haveItemsQueued = function () {
return this._isTickUsed || this._haveDrainedQueues;
2015-06-08 06:38:56 +00:00
};
2016-01-09 02:07:10 +00:00
Async.prototype.fatalError = function(e, isNode) {
if (isNode) {
process.stderr.write("Fatal " + (e instanceof Error ? e.stack : e));
process.exit(2);
} else {
this.throwLater(e);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Async.prototype.throwLater = function(fn, arg) {
if (arguments.length === 1) {
arg = fn;
fn = function () { throw arg; };
}
if (typeof setTimeout !== "undefined") {
setTimeout(function() {
fn(arg);
}, 0);
} else try {
this._schedule(function() {
fn(arg);
});
} catch (e) {
2016-01-09 02:07:10 +00:00
throw new Error("No async scheduler available\u000a\u000a See http://goo.gl/MqrFmX\u000a");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
2015-06-08 06:38:56 +00:00
function AsyncInvokeLater(fn, receiver, arg) {
this._lateQueue.push(fn, receiver, arg);
this._queueTick();
2015-06-08 06:38:56 +00:00
}
2015-06-08 06:38:56 +00:00
function AsyncInvoke(fn, receiver, arg) {
this._normalQueue.push(fn, receiver, arg);
this._queueTick();
2015-06-08 06:38:56 +00:00
}
2015-06-08 06:38:56 +00:00
function AsyncSettlePromises(promise) {
2015-10-29 06:23:22 +00:00
this._normalQueue._pushOne(promise);
this._queueTick();
2015-06-08 06:38:56 +00:00
}
2015-06-08 06:38:56 +00:00
if (!util.hasDevTools) {
Async.prototype.invokeLater = AsyncInvokeLater;
Async.prototype.invoke = AsyncInvoke;
Async.prototype.settlePromises = AsyncSettlePromises;
} else {
Async.prototype.invokeLater = function (fn, receiver, arg) {
if (this._trampolineEnabled) {
AsyncInvokeLater.call(this, fn, receiver, arg);
} else {
2015-10-29 06:23:22 +00:00
this._schedule(function() {
setTimeout(function() {
fn.call(receiver, arg);
}, 100);
});
2015-06-08 06:38:56 +00:00
}
};
Async.prototype.invoke = function (fn, receiver, arg) {
if (this._trampolineEnabled) {
AsyncInvoke.call(this, fn, receiver, arg);
} else {
2015-10-29 06:23:22 +00:00
this._schedule(function() {
2015-06-08 06:38:56 +00:00
fn.call(receiver, arg);
2015-10-29 06:23:22 +00:00
});
2015-06-08 06:38:56 +00:00
}
};
Async.prototype.settlePromises = function(promise) {
if (this._trampolineEnabled) {
AsyncSettlePromises.call(this, promise);
} else {
2015-10-29 06:23:22 +00:00
this._schedule(function() {
2015-06-08 06:38:56 +00:00
promise._settlePromises();
2015-10-29 06:23:22 +00:00
});
2015-06-08 06:38:56 +00:00
}
};
}
Async.prototype.invokeFirst = function (fn, receiver, arg) {
this._normalQueue.unshift(fn, receiver, arg);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._queueTick();
};
Async.prototype._drainQueue = function(queue) {
while (queue.length() > 0) {
var fn = queue.shift();
if (typeof fn !== "function") {
fn._settlePromises();
continue;
}
var receiver = queue.shift();
var arg = queue.shift();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
fn.call(receiver, arg);
}
};
Async.prototype._drainQueues = function () {
this._drainQueue(this._normalQueue);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._reset();
2016-01-09 02:07:10 +00:00
this._haveDrainedQueues = true;
this._drainQueue(this._lateQueue);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Async.prototype._queueTick = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!this._isTickUsed) {
this._isTickUsed = true;
this._schedule(this.drainQueues);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
Async.prototype._reset = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._isTickUsed = false;
};
2016-01-09 02:07:10 +00:00
module.exports = Async;
module.exports.firstLineError = firstLineError;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
},{"./queue":26,"./schedule":29,"./util":36}],3:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
2016-01-09 02:07:10 +00:00
module.exports = function(Promise, INTERNAL, tryConvertToPromise, debug) {
var calledBind = false;
2015-06-08 06:38:56 +00:00
var rejectThis = function(_, e) {
this._reject(e);
};
2015-06-08 06:38:56 +00:00
var targetRejected = function(e, context) {
context.promiseRejectionQueued = true;
context.bindingPromise._then(rejectThis, rejectThis, null, this, e);
};
2015-06-08 06:38:56 +00:00
var bindingResolved = function(thisArg, context) {
2016-01-09 02:07:10 +00:00
if (((this._bitField & 50397184) === 0)) {
2015-06-08 06:38:56 +00:00
this._resolveCallback(context.target);
}
};
2015-06-08 06:38:56 +00:00
var bindingRejected = function(e, context) {
if (!context.promiseRejectionQueued) this._reject(e);
};
2015-06-08 06:38:56 +00:00
Promise.prototype.bind = function (thisArg) {
2016-01-09 02:07:10 +00:00
if (!calledBind) {
calledBind = true;
Promise.prototype._propagateFrom = debug.propagateFromFunction();
Promise.prototype._boundValue = debug.boundValueFunction();
}
2015-06-08 06:38:56 +00:00
var maybePromise = tryConvertToPromise(thisArg);
var ret = new Promise(INTERNAL);
ret._propagateFrom(this, 1);
var target = this._target();
2015-10-29 06:23:22 +00:00
ret._setBoundTo(maybePromise);
2015-06-08 06:38:56 +00:00
if (maybePromise instanceof Promise) {
var context = {
promiseRejectionQueued: false,
promise: ret,
target: target,
bindingPromise: maybePromise
};
2016-01-09 02:07:10 +00:00
target._then(INTERNAL, targetRejected, undefined, ret, context);
2015-06-08 06:38:56 +00:00
maybePromise._then(
2016-01-09 02:07:10 +00:00
bindingResolved, bindingRejected, undefined, ret, context);
ret._setOnCancel(maybePromise);
2015-06-08 06:38:56 +00:00
} else {
ret._resolveCallback(target);
}
return ret;
};
Promise.prototype._setBoundTo = function (obj) {
2015-06-08 06:38:56 +00:00
if (obj !== undefined) {
2016-01-09 02:07:10 +00:00
this._bitField = this._bitField | 2097152;
2015-06-08 06:38:56 +00:00
this._boundTo = obj;
} else {
2016-01-09 02:07:10 +00:00
this._bitField = this._bitField & (~2097152);
2015-06-08 06:38:56 +00:00
}
};
Promise.prototype._isBound = function () {
2016-01-09 02:07:10 +00:00
return (this._bitField & 2097152) === 2097152;
2015-06-08 06:38:56 +00:00
};
Promise.bind = function (thisArg, value) {
2016-01-09 02:07:10 +00:00
return Promise.resolve(value).bind(thisArg);
};
};
},{}],4:[function(_dereq_,module,exports){
"use strict";
var old;
if (typeof Promise !== "undefined") old = Promise;
function noConflict() {
try { if (Promise === bluebird) Promise = old; }
catch (e) {}
return bluebird;
}
2016-01-09 02:07:10 +00:00
var bluebird = _dereq_("./promise")();
bluebird.noConflict = noConflict;
module.exports = bluebird;
2016-01-09 02:07:10 +00:00
},{"./promise":22}],5:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
var cr = Object.create;
if (cr) {
var callerCache = cr(null);
var getterCache = cr(null);
callerCache[" size"] = getterCache[" size"] = 0;
}
module.exports = function(Promise) {
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var canEvaluate = util.canEvaluate;
var isIdentifier = util.isIdentifier;
var getMethodCaller;
var getGetter;
if (!true) {
var makeMethodCaller = function (methodName) {
2015-02-03 20:28:54 +00:00
return new Function("ensureMethod", " \n\
return function(obj) { \n\
'use strict' \n\
var len = this.length; \n\
ensureMethod(obj, 'methodName'); \n\
switch(len) { \n\
case 1: return obj.methodName(this[0]); \n\
case 2: return obj.methodName(this[0], this[1]); \n\
case 3: return obj.methodName(this[0], this[1], this[2]); \n\
case 0: return obj.methodName(); \n\
default: \n\
return obj.methodName.apply(obj, this); \n\
} \n\
}; \n\
".replace(/methodName/g, methodName))(ensureMethod);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var makeGetter = function (propertyName) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return new Function("obj", " \n\
'use strict'; \n\
return obj.propertyName; \n\
".replace("propertyName", propertyName));
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var getCompiled = function(name, compiler, cache) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = cache[name];
if (typeof ret !== "function") {
if (!isIdentifier(name)) {
return null;
}
ret = compiler(name);
cache[name] = ret;
cache[" size"]++;
if (cache[" size"] > 512) {
var keys = Object.keys(cache);
for (var i = 0; i < 256; ++i) delete cache[keys[i]];
cache[" size"] = keys.length - 256;
}
}
return ret;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
getMethodCaller = function(name) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return getCompiled(name, makeMethodCaller, callerCache);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
getGetter = function(name) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return getCompiled(name, makeGetter, getterCache);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2015-02-03 20:28:54 +00:00
function ensureMethod(obj, methodName) {
var fn;
if (obj != null) fn = obj[methodName];
if (typeof fn !== "function") {
var message = "Object " + util.classString(obj) + " has no method '" +
util.toString(methodName) + "'";
throw new Promise.TypeError(message);
}
return fn;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function caller(obj) {
2015-02-03 20:28:54 +00:00
var methodName = this.pop();
var fn = ensureMethod(obj, methodName);
return fn.apply(obj, this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
Promise.prototype.call = function (methodName) {
2016-01-09 02:07:10 +00:00
var args = [].slice.call(arguments, 1);;
if (!true) {
if (canEvaluate) {
var maybeCaller = getMethodCaller(methodName);
if (maybeCaller !== null) {
return this._then(
maybeCaller, undefined, undefined, args, undefined);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
args.push(methodName);
return this._then(caller, undefined, undefined, args, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
function namedGetter(obj) {
return obj[this];
}
function indexedGetter(obj) {
var index = +this;
if (index < 0) index = Math.max(0, index + obj.length);
return obj[index];
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
Promise.prototype.get = function (propertyName) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var isIndex = (typeof propertyName === "number");
var getter;
if (!isIndex) {
if (canEvaluate) {
var maybeGetter = getGetter(propertyName);
getter = maybeGetter !== null ? maybeGetter : namedGetter;
} else {
getter = namedGetter;
}
} else {
getter = indexedGetter;
}
return this._then(getter, undefined, undefined, propertyName, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],6:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
2016-01-09 02:07:10 +00:00
module.exports = function(Promise, PromiseArray, apiRejection, debug) {
var util = _dereq_("./util");
var tryCatch = util.tryCatch;
var errorObj = util.errorObj;
var async = Promise._async;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
Promise.prototype["break"] = Promise.prototype.cancel = function() {
if (!debug.cancellation()) return this._warn("cancellation is disabled");
var promise = this;
var child = promise;
while (promise.isCancellable()) {
if (!promise._cancelBy(child)) {
if (child._isFollowing()) {
child._followee().cancel();
} else {
child._cancelBranched();
}
break;
}
var parent = promise._cancellationParent;
if (parent == null || !parent.isCancellable()) {
if (promise._isFollowing()) {
promise._followee().cancel();
} else {
promise._cancelBranched();
}
break;
} else {
if (promise._isFollowing()) promise._followee().cancel();
child = promise;
promise = parent;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
2016-01-09 02:07:10 +00:00
Promise.prototype._branchHasCancelled = function() {
this._branchesRemainingToCancel--;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._enoughBranchesHaveCancelled = function() {
return this._branchesRemainingToCancel === undefined ||
this._branchesRemainingToCancel <= 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._cancelBy = function(canceller) {
if (canceller === this) {
this._branchesRemainingToCancel = 0;
this._invokeOnCancel();
return true;
} else {
this._branchHasCancelled();
if (this._enoughBranchesHaveCancelled()) {
this._invokeOnCancel();
return true;
}
}
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._cancelBranched = function() {
if (this._enoughBranchesHaveCancelled()) {
this._cancel();
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._cancel = function() {
if (!this.isCancellable()) return;
this._setCancelled();
async.invoke(this._cancelPromises, this, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._cancelPromises = function() {
if (this._length() > 0) this._settlePromises();
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
Promise.prototype._unsetOnCancel = function() {
this._onCancelField = undefined;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
Promise.prototype.isCancellable = function() {
return this.isPending() && !this.isCancelled();
};
2016-01-09 02:07:10 +00:00
Promise.prototype._doInvokeOnCancel = function(onCancelCallback, internalOnly) {
if (util.isArray(onCancelCallback)) {
for (var i = 0; i < onCancelCallback.length; ++i) {
this._doInvokeOnCancel(onCancelCallback[i], internalOnly);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
} else if (onCancelCallback !== undefined) {
if (typeof onCancelCallback === "function") {
if (!internalOnly) {
var e = tryCatch(onCancelCallback).call(this._boundValue());
if (e === errorObj) {
this._attachExtraTrace(e.e);
async.throwLater(e.e);
}
}
2016-01-09 02:07:10 +00:00
} else {
onCancelCallback._resultCancelled(this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
Promise.prototype._invokeOnCancel = function() {
var onCancelCallback = this._onCancel();
this._unsetOnCancel();
async.invoke(this._doInvokeOnCancel, this, onCancelCallback);
};
2016-01-09 02:07:10 +00:00
Promise.prototype._invokeInternalOnCancel = function() {
if (this.isCancellable()) {
this._doInvokeOnCancel(this._onCancel(), true);
this._unsetOnCancel();
}
};
2016-01-09 02:07:10 +00:00
Promise.prototype._resultCancelled = function() {
this.cancel();
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],7:[function(_dereq_,module,exports){
"use strict";
module.exports = function(NEXT_FILTER) {
var util = _dereq_("./util");
var getKeys = _dereq_("./es5").keys;
var tryCatch = util.tryCatch;
var errorObj = util.errorObj;
2016-01-09 02:07:10 +00:00
function catchFilter(instances, cb, promise) {
return function(e) {
var boundTo = promise._boundValue();
predicateLoop: for (var i = 0; i < instances.length; ++i) {
var item = instances[i];
if (item === Error ||
(item != null && item.prototype instanceof Error)) {
if (e instanceof item) {
return tryCatch(cb).call(boundTo, e);
}
} else if (typeof item === "function") {
var matchesPredicate = tryCatch(item).call(boundTo, e);
if (matchesPredicate === errorObj) {
return matchesPredicate;
} else if (matchesPredicate) {
return tryCatch(cb).call(boundTo, e);
}
} else if (util.isObject(e)) {
var keys = getKeys(item);
for (var j = 0; j < keys.length; ++j) {
var key = keys[j];
if (item[key] != e[key]) {
continue predicateLoop;
}
}
return tryCatch(cb).call(boundTo, e);
}
}
2016-01-09 02:07:10 +00:00
return NEXT_FILTER;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return catchFilter;
};
2016-01-09 02:07:10 +00:00
},{"./es5":13,"./util":36}],8:[function(_dereq_,module,exports){
"use strict";
module.exports = function(Promise) {
var longStackTraces = false;
var contextStack = [];
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
Promise.prototype._promiseCreated = function() {};
Promise.prototype._pushContext = function() {};
Promise.prototype._popContext = function() {return null;};
Promise._peekContext = Promise.prototype._peekContext = function() {};
function Context() {
this._trace = new Context.CapturedTrace(peekContext());
}
2016-01-09 02:07:10 +00:00
Context.prototype._pushContext = function () {
if (this._trace !== undefined) {
this._trace._promiseCreated = null;
contextStack.push(this._trace);
}
};
2016-01-09 02:07:10 +00:00
Context.prototype._popContext = function () {
if (this._trace !== undefined) {
var trace = contextStack.pop();
var ret = trace._promiseCreated;
trace._promiseCreated = null;
return ret;
}
2016-01-09 02:07:10 +00:00
return null;
};
function createContext() {
if (longStackTraces) return new Context();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
function peekContext() {
var lastIndex = contextStack.length - 1;
if (lastIndex >= 0) {
return contextStack[lastIndex];
}
2016-01-09 02:07:10 +00:00
return undefined;
}
2016-01-09 02:07:10 +00:00
Context.CapturedTrace = null;
Context.create = createContext;
Context.deactivateLongStackTraces = function() {};
Context.activateLongStackTraces = function() {
var Promise_pushContext = Promise.prototype._pushContext;
var Promise_popContext = Promise.prototype._popContext;
var Promise_PeekContext = Promise._peekContext;
var Promise_peekContext = Promise.prototype._peekContext;
var Promise_promiseCreated = Promise.prototype._promiseCreated;
Context.deactivateLongStackTraces = function() {
Promise.prototype._pushContext = Promise_pushContext;
Promise.prototype._popContext = Promise_popContext;
Promise._peekContext = Promise_PeekContext;
Promise.prototype._peekContext = Promise_peekContext;
Promise.prototype._promiseCreated = Promise_promiseCreated;
longStackTraces = false;
};
longStackTraces = true;
Promise.prototype._pushContext = Context.prototype._pushContext;
Promise.prototype._popContext = Context.prototype._popContext;
Promise._peekContext = Promise.prototype._peekContext = peekContext;
Promise.prototype._promiseCreated = function() {
var ctx = this._peekContext();
if (ctx && ctx._promiseCreated == null) ctx._promiseCreated = this;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
return Context;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
},{}],9:[function(_dereq_,module,exports){
"use strict";
module.exports = function(Promise, Context) {
var getDomain = Promise._getDomain;
var async = Promise._async;
var Warning = _dereq_("./errors").Warning;
var util = _dereq_("./util");
var canAttachTrace = util.canAttachTrace;
var unhandledRejectionHandled;
var possiblyUnhandledRejection;
var bluebirdFramePattern =
/[\\\/]bluebird[\\\/]js[\\\/](release|debug|instrumented)/;
var stackFramePattern = null;
var formatStack = null;
var indentStackFrames = false;
var printWarning;
var debugging = !!(util.env("BLUEBIRD_DEBUG") != 0 &&
(true ||
util.env("BLUEBIRD_DEBUG") ||
util.env("NODE_ENV") === "development"));
var warnings = !!(util.env("BLUEBIRD_WARNINGS") != 0 &&
(debugging || util.env("BLUEBIRD_WARNINGS")));
var longStackTraces = !!(util.env("BLUEBIRD_LONG_STACK_TRACES") != 0 &&
(debugging || util.env("BLUEBIRD_LONG_STACK_TRACES")));
var wForgottenReturn = util.env("BLUEBIRD_W_FORGOTTEN_RETURN") != 0 &&
(warnings || !!util.env("BLUEBIRD_W_FORGOTTEN_RETURN"));
Promise.prototype.suppressUnhandledRejections = function() {
var target = this._target();
target._bitField = ((target._bitField & (~1048576)) |
524288);
};
Promise.prototype._ensurePossibleRejectionHandled = function () {
if ((this._bitField & 524288) !== 0) return;
this._setRejectionIsUnhandled();
async.invokeLater(this._notifyUnhandledRejection, this, undefined);
};
Promise.prototype._notifyUnhandledRejectionIsHandled = function () {
fireRejectionEvent("rejectionHandled",
unhandledRejectionHandled, undefined, this);
};
Promise.prototype._setReturnedNonUndefined = function() {
this._bitField = this._bitField | 268435456;
};
Promise.prototype._returnedNonUndefined = function() {
return (this._bitField & 268435456) !== 0;
};
Promise.prototype._notifyUnhandledRejection = function () {
if (this._isRejectionUnhandled()) {
var reason = this._settledValue();
this._setUnhandledRejectionIsNotified();
fireRejectionEvent("unhandledRejection",
possiblyUnhandledRejection, reason, this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
2016-01-09 02:07:10 +00:00
Promise.prototype._setUnhandledRejectionIsNotified = function () {
this._bitField = this._bitField | 262144;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
Promise.prototype._unsetUnhandledRejectionIsNotified = function () {
this._bitField = this._bitField & (~262144);
};
Promise.prototype._isUnhandledRejectionNotified = function () {
return (this._bitField & 262144) > 0;
};
Promise.prototype._setRejectionIsUnhandled = function () {
this._bitField = this._bitField | 1048576;
};
Promise.prototype._unsetRejectionIsUnhandled = function () {
this._bitField = this._bitField & (~1048576);
if (this._isUnhandledRejectionNotified()) {
this._unsetUnhandledRejectionIsNotified();
this._notifyUnhandledRejectionIsHandled();
}
};
Promise.prototype._isRejectionUnhandled = function () {
return (this._bitField & 1048576) > 0;
};
Promise.prototype._warn = function(message, shouldUseOwnTrace, promise) {
return warn(message, shouldUseOwnTrace, promise || this);
};
Promise.onPossiblyUnhandledRejection = function (fn) {
var domain = getDomain();
possiblyUnhandledRejection =
typeof fn === "function" ? (domain === null ? fn : domain.bind(fn))
: undefined;
};
Promise.onUnhandledRejectionHandled = function (fn) {
var domain = getDomain();
unhandledRejectionHandled =
typeof fn === "function" ? (domain === null ? fn : domain.bind(fn))
: undefined;
};
var disableLongStackTraces = function() {};
Promise.longStackTraces = function () {
if (async.haveItemsQueued() && !config.longStackTraces) {
throw new Error("cannot enable long stack traces after promises have been created\u000a\u000a See http://goo.gl/MqrFmX\u000a");
}
if (!config.longStackTraces && longStackTracesIsSupported()) {
var Promise_captureStackTrace = Promise.prototype._captureStackTrace;
var Promise_attachExtraTrace = Promise.prototype._attachExtraTrace;
config.longStackTraces = true;
disableLongStackTraces = function() {
if (async.haveItemsQueued() && !config.longStackTraces) {
throw new Error("cannot enable long stack traces after promises have been created\u000a\u000a See http://goo.gl/MqrFmX\u000a");
}
Promise.prototype._captureStackTrace = Promise_captureStackTrace;
Promise.prototype._attachExtraTrace = Promise_attachExtraTrace;
Context.deactivateLongStackTraces();
async.enableTrampoline();
config.longStackTraces = false;
};
Promise.prototype._captureStackTrace = longStackTracesCaptureStackTrace;
Promise.prototype._attachExtraTrace = longStackTracesAttachExtraTrace;
Context.activateLongStackTraces();
async.disableTrampolineIfNecessary();
}
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
Promise.hasLongStackTraces = function () {
return config.longStackTraces && longStackTracesIsSupported();
};
Promise.config = function(opts) {
opts = Object(opts);
if ("longStackTraces" in opts) {
if (opts.longStackTraces) {
Promise.longStackTraces();
} else if (!opts.longStackTraces && Promise.hasLongStackTraces()) {
disableLongStackTraces();
}
}
if ("warnings" in opts) {
var warningsOption = opts.warnings;
config.warnings = !!warningsOption;
wForgottenReturn = config.warnings;
if (util.isObject(warningsOption)) {
if ("wForgottenReturn" in warningsOption) {
wForgottenReturn = !!warningsOption.wForgottenReturn;
}
}
}
if ("cancellation" in opts && opts.cancellation && !config.cancellation) {
if (async.haveItemsQueued()) {
throw new Error(
"cannot enable cancellation after promises are in use");
}
Promise.prototype._clearCancellationData =
cancellationClearCancellationData;
Promise.prototype._propagateFrom = cancellationPropagateFrom;
Promise.prototype._onCancel = cancellationOnCancel;
Promise.prototype._setOnCancel = cancellationSetOnCancel;
Promise.prototype._attachCancellationCallback =
cancellationAttachCancellationCallback;
Promise.prototype._execute = cancellationExecute;
propagateFromFunction = cancellationPropagateFrom;
config.cancellation = true;
}
};
Promise.prototype._execute = function(executor, resolve, reject) {
try {
executor(resolve, reject);
} catch (e) {
return e;
}
};
Promise.prototype._onCancel = function () {};
Promise.prototype._setOnCancel = function (handler) { ; };
Promise.prototype._attachCancellationCallback = function(onCancel) {
;
};
Promise.prototype._captureStackTrace = function () {};
Promise.prototype._attachExtraTrace = function () {};
Promise.prototype._clearCancellationData = function() {};
Promise.prototype._propagateFrom = function (parent, flags) {
;
;
};
function cancellationExecute(executor, resolve, reject) {
var promise = this;
try {
executor(resolve, reject, function(onCancel) {
if (typeof onCancel !== "function") {
throw new TypeError("onCancel must be a function, got: " +
util.toString(onCancel));
}
promise._attachCancellationCallback(onCancel);
});
} catch (e) {
return e;
}
}
function cancellationAttachCancellationCallback(onCancel) {
if (!this.isCancellable()) return this;
var previousOnCancel = this._onCancel();
if (previousOnCancel !== undefined) {
if (util.isArray(previousOnCancel)) {
previousOnCancel.push(onCancel);
} else {
this._setOnCancel([previousOnCancel, onCancel]);
}
} else {
this._setOnCancel(onCancel);
}
}
function cancellationOnCancel() {
return this._onCancelField;
}
function cancellationSetOnCancel(onCancel) {
this._onCancelField = onCancel;
}
function cancellationClearCancellationData() {
this._cancellationParent = undefined;
this._onCancelField = undefined;
}
function cancellationPropagateFrom(parent, flags) {
if ((flags & 1) !== 0) {
this._cancellationParent = parent;
var branchesRemainingToCancel = parent._branchesRemainingToCancel;
if (branchesRemainingToCancel === undefined) {
branchesRemainingToCancel = 0;
}
parent._branchesRemainingToCancel = branchesRemainingToCancel + 1;
}
if ((flags & 2) !== 0 && parent._isBound()) {
this._setBoundTo(parent._boundTo);
}
}
function bindingPropagateFrom(parent, flags) {
if ((flags & 2) !== 0 && parent._isBound()) {
this._setBoundTo(parent._boundTo);
}
}
var propagateFromFunction = bindingPropagateFrom;
function boundValueFunction() {
var ret = this._boundTo;
if (ret !== undefined) {
if (ret instanceof Promise) {
if (ret.isFulfilled()) {
return ret.value();
} else {
return undefined;
}
}
}
return ret;
}
function longStackTracesCaptureStackTrace() {
this._trace = new CapturedTrace(this._peekContext());
}
function longStackTracesAttachExtraTrace(error, ignoreSelf) {
if (canAttachTrace(error)) {
var trace = this._trace;
if (trace !== undefined) {
if (ignoreSelf) trace = trace._parent;
}
if (trace !== undefined) {
trace.attachExtraTrace(error);
} else if (!error.__stackCleaned__) {
var parsed = parseStackAndMessage(error);
util.notEnumerableProp(error, "stack",
parsed.message + "\n" + parsed.stack.join("\n"));
util.notEnumerableProp(error, "__stackCleaned__", true);
}
}
}
function checkForgottenReturns(returnValue, promiseCreated, name, promise,
parent) {
if (returnValue === undefined && promiseCreated !== null &&
wForgottenReturn) {
if (parent !== undefined && parent._returnedNonUndefined()) return;
if (name) name = name + " ";
var msg = "a promise was created in a " + name +
"handler but was not returned from it";
promise._warn(msg, true, promiseCreated);
}
}
function deprecated(name, replacement) {
var message = name +
" is deprecated and will be removed in a future version.";
if (replacement) message += " Use " + replacement + " instead.";
return warn(message);
}
function warn(message, shouldUseOwnTrace, promise) {
if (!config.warnings) return;
var warning = new Warning(message);
var ctx;
if (shouldUseOwnTrace) {
promise._attachExtraTrace(warning);
} else if (config.longStackTraces && (ctx = Promise._peekContext())) {
ctx.attachExtraTrace(warning);
} else {
var parsed = parseStackAndMessage(warning);
warning.stack = parsed.message + "\n" + parsed.stack.join("\n");
}
formatAndLogError(warning, "", true);
}
function reconstructStack(message, stacks) {
for (var i = 0; i < stacks.length - 1; ++i) {
stacks[i].push("From previous event:");
stacks[i] = stacks[i].join("\n");
}
if (i < stacks.length) {
stacks[i] = stacks[i].join("\n");
}
return message + "\n" + stacks.join("\n");
}
function removeDuplicateOrEmptyJumps(stacks) {
for (var i = 0; i < stacks.length; ++i) {
if (stacks[i].length === 0 ||
((i + 1 < stacks.length) && stacks[i][0] === stacks[i+1][0])) {
stacks.splice(i, 1);
i--;
}
}
}
function removeCommonRoots(stacks) {
var current = stacks[0];
for (var i = 1; i < stacks.length; ++i) {
var prev = stacks[i];
var currentLastIndex = current.length - 1;
var currentLastLine = current[currentLastIndex];
var commonRootMeetPoint = -1;
for (var j = prev.length - 1; j >= 0; --j) {
if (prev[j] === currentLastLine) {
commonRootMeetPoint = j;
break;
}
}
for (var j = commonRootMeetPoint; j >= 0; --j) {
var line = prev[j];
if (current[currentLastIndex] === line) {
current.pop();
currentLastIndex--;
} else {
break;
}
}
current = prev;
}
}
function cleanStack(stack) {
var ret = [];
for (var i = 0; i < stack.length; ++i) {
var line = stack[i];
var isTraceLine = " (No stack trace)" === line ||
stackFramePattern.test(line);
var isInternalFrame = isTraceLine && shouldIgnore(line);
if (isTraceLine && !isInternalFrame) {
if (indentStackFrames && line.charAt(0) !== " ") {
line = " " + line;
}
ret.push(line);
}
}
return ret;
}
function stackFramesAsArray(error) {
var stack = error.stack.replace(/\s+$/g, "").split("\n");
for (var i = 0; i < stack.length; ++i) {
var line = stack[i];
if (" (No stack trace)" === line || stackFramePattern.test(line)) {
break;
}
}
if (i > 0) {
stack = stack.slice(i);
}
return stack;
}
function parseStackAndMessage(error) {
var stack = error.stack;
var message = error.toString();
stack = typeof stack === "string" && stack.length > 0
? stackFramesAsArray(error) : [" (No stack trace)"];
return {
message: message,
stack: cleanStack(stack)
};
}
function formatAndLogError(error, title, isSoft) {
if (typeof console !== "undefined") {
var message;
if (util.isObject(error)) {
var stack = error.stack;
message = title + formatStack(stack, error);
} else {
message = title + String(error);
}
if (typeof printWarning === "function") {
printWarning(message, isSoft);
} else if (typeof console.log === "function" ||
typeof console.log === "object") {
console.log(message);
}
}
}
function fireRejectionEvent(name, localHandler, reason, promise) {
var localEventFired = false;
try {
if (typeof localHandler === "function") {
localEventFired = true;
if (name === "rejectionHandled") {
localHandler(promise);
} else {
localHandler(reason, promise);
}
}
} catch (e) {
async.throwLater(e);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var globalEventFired = false;
try {
globalEventFired = fireGlobalEvent(name, reason, promise);
} catch (e) {
globalEventFired = true;
async.throwLater(e);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var domEventFired = false;
if (fireDomEvent) {
try {
domEventFired = fireDomEvent(name.toLowerCase(), {
reason: reason,
promise: promise
});
} catch (e) {
domEventFired = true;
async.throwLater(e);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
if (!globalEventFired && !localEventFired && !domEventFired &&
name === "unhandledRejection") {
2016-01-09 02:07:10 +00:00
formatAndLogError(reason, "Unhandled rejection ");
}
2016-01-09 02:07:10 +00:00
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function formatNonError(obj) {
var str;
if (typeof obj === "function") {
str = "[function " +
(obj.name || "anonymous") +
"]";
} else {
2016-01-09 02:07:10 +00:00
str = obj && typeof obj.toString === "function"
? obj.toString() : util.toString(obj);
var ruselessToString = /\[object [a-zA-Z0-9$_]+\]/;
if (ruselessToString.test(str)) {
try {
var newStr = JSON.stringify(obj);
str = newStr;
}
catch(e) {
}
}
if (str.length === 0) {
str = "(empty array)";
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
return ("(<" + snip(str) + ">, no stack trace)");
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function snip(str) {
var maxChars = 41;
if (str.length < maxChars) {
return str;
}
return str.substr(0, maxChars - 3) + "...";
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
function longStackTracesIsSupported() {
return typeof captureStackTrace === "function";
}
var shouldIgnore = function() { return false; };
var parseLineInfoRegex = /[\/<\(]([^:\/]+):(\d+):(?:\d+)\)?\s*$/;
function parseLineInfo(line) {
var matches = line.match(parseLineInfoRegex);
if (matches) {
return {
fileName: matches[1],
line: parseInt(matches[2], 10)
};
}
}
2016-01-09 02:07:10 +00:00
function setBounds(firstLineError, lastLineError) {
if (!longStackTracesIsSupported()) return;
var firstStackLines = firstLineError.stack.split("\n");
var lastStackLines = lastLineError.stack.split("\n");
var firstIndex = -1;
var lastIndex = -1;
var firstFileName;
var lastFileName;
for (var i = 0; i < firstStackLines.length; ++i) {
var result = parseLineInfo(firstStackLines[i]);
if (result) {
firstFileName = result.fileName;
firstIndex = result.line;
break;
}
}
for (var i = 0; i < lastStackLines.length; ++i) {
var result = parseLineInfo(lastStackLines[i]);
if (result) {
lastFileName = result.fileName;
lastIndex = result.line;
break;
}
}
if (firstIndex < 0 || lastIndex < 0 || !firstFileName || !lastFileName ||
firstFileName !== lastFileName || firstIndex >= lastIndex) {
return;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
shouldIgnore = function(line) {
if (bluebirdFramePattern.test(line)) return true;
var info = parseLineInfo(line);
if (info) {
if (info.fileName === firstFileName &&
(firstIndex <= info.line && info.line <= lastIndex)) {
return true;
}
}
return false;
};
2016-01-09 02:07:10 +00:00
}
function CapturedTrace(parent) {
this._parent = parent;
this._promisesCreated = 0;
var length = this._length = 1 + (parent === undefined ? 0 : parent._length);
captureStackTrace(this, CapturedTrace);
if (length > 32) this.uncycle();
}
util.inherits(CapturedTrace, Error);
Context.CapturedTrace = CapturedTrace;
CapturedTrace.prototype.uncycle = function() {
var length = this._length;
if (length < 2) return;
var nodes = [];
var stackToIndex = {};
for (var i = 0, node = this; node !== undefined; ++i) {
nodes.push(node);
node = node._parent;
}
length = this._length = i;
for (var i = length - 1; i >= 0; --i) {
var stack = nodes[i].stack;
if (stackToIndex[stack] === undefined) {
stackToIndex[stack] = i;
}
}
for (var i = 0; i < length; ++i) {
var currentStack = nodes[i].stack;
var index = stackToIndex[currentStack];
if (index !== undefined && index !== i) {
if (index > 0) {
nodes[index - 1]._parent = undefined;
nodes[index - 1]._length = 1;
}
nodes[i]._parent = undefined;
nodes[i]._length = 1;
var cycleEdgeNode = i > 0 ? nodes[i - 1] : this;
if (index < length - 1) {
cycleEdgeNode._parent = nodes[index + 1];
cycleEdgeNode._parent.uncycle();
cycleEdgeNode._length =
cycleEdgeNode._parent._length + 1;
} else {
cycleEdgeNode._parent = undefined;
cycleEdgeNode._length = 1;
}
var currentChildLength = cycleEdgeNode._length + 1;
for (var j = i - 2; j >= 0; --j) {
nodes[j]._length = currentChildLength;
currentChildLength++;
}
return;
}
}
};
CapturedTrace.prototype.attachExtraTrace = function(error) {
if (error.__stackCleaned__) return;
this.uncycle();
var parsed = parseStackAndMessage(error);
var message = parsed.message;
var stacks = [parsed.stack];
var trace = this;
while (trace !== undefined) {
stacks.push(cleanStack(trace.stack.split("\n")));
trace = trace._parent;
}
removeCommonRoots(stacks);
removeDuplicateOrEmptyJumps(stacks);
util.notEnumerableProp(error, "stack", reconstructStack(message, stacks));
util.notEnumerableProp(error, "__stackCleaned__", true);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
var captureStackTrace = (function stackDetection() {
var v8stackFramePattern = /^\s*at\s*/;
var v8stackFormatter = function(stack, error) {
if (typeof stack === "string") return stack;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (error.name !== undefined &&
error.message !== undefined) {
return error.toString();
}
return formatNonError(error);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof Error.stackTraceLimit === "number" &&
typeof Error.captureStackTrace === "function") {
2016-01-09 02:07:10 +00:00
Error.stackTraceLimit += 6;
stackFramePattern = v8stackFramePattern;
formatStack = v8stackFormatter;
var captureStackTrace = Error.captureStackTrace;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
shouldIgnore = function(line) {
return bluebirdFramePattern.test(line);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
return function(receiver, ignoreUntil) {
2016-01-09 02:07:10 +00:00
Error.stackTraceLimit += 6;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
captureStackTrace(receiver, ignoreUntil);
2016-01-09 02:07:10 +00:00
Error.stackTraceLimit -= 6;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
}
var err = new Error();
if (typeof err.stack === "string" &&
err.stack.split("\n")[0].indexOf("stackDetection@") >= 0) {
stackFramePattern = /@/;
formatStack = v8stackFormatter;
indentStackFrames = true;
return function captureStackTrace(o) {
o.stack = new Error().stack;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var hasStackAfterThrow;
try { throw new Error(); }
catch(e) {
hasStackAfterThrow = ("stack" in e);
}
2015-10-29 06:23:22 +00:00
if (!("stack" in err) && hasStackAfterThrow &&
typeof Error.stackTraceLimit === "number") {
stackFramePattern = v8stackFramePattern;
formatStack = v8stackFormatter;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return function captureStackTrace(o) {
2016-01-09 02:07:10 +00:00
Error.stackTraceLimit += 6;
try { throw new Error(); }
catch(e) { o.stack = e.stack; }
2016-01-09 02:07:10 +00:00
Error.stackTraceLimit -= 6;
};
}
formatStack = function(stack, error) {
if (typeof stack === "string") return stack;
if ((typeof error === "object" ||
typeof error === "function") &&
error.name !== undefined &&
error.message !== undefined) {
return error.toString();
}
return formatNonError(error);
};
return null;
})([]);
var fireDomEvent;
var fireGlobalEvent = (function() {
2015-02-03 20:28:54 +00:00
if (util.isNode) {
return function(name, reason, promise) {
if (name === "rejectionHandled") {
return process.emit(name, promise);
} else {
return process.emit(name, reason, promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
} else {
2016-01-09 02:07:10 +00:00
var globalObject = typeof self !== "undefined" ? self :
typeof window !== "undefined" ? window :
typeof global !== "undefined" ? global :
this !== undefined ? this : null;
if (!globalObject) {
return function() {
return false;
};
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
try {
var event = document.createEvent("CustomEvent");
event.initCustomEvent("testingtheevent", false, true, {});
globalObject.dispatchEvent(event);
fireDomEvent = function(type, detail) {
var event = document.createEvent("CustomEvent");
event.initCustomEvent(type, false, true, detail);
return !globalObject.dispatchEvent(event);
};
2016-01-09 02:07:10 +00:00
} catch (e) {}
var toWindowMethodNameMap = {};
toWindowMethodNameMap["unhandledRejection"] = ("on" +
"unhandledRejection").toLowerCase();
toWindowMethodNameMap["rejectionHandled"] = ("on" +
"rejectionHandled").toLowerCase();
return function(name, reason, promise) {
var methodName = toWindowMethodNameMap[name];
2016-01-09 02:07:10 +00:00
var method = globalObject[methodName];
if (!method) return false;
if (name === "rejectionHandled") {
2016-01-09 02:07:10 +00:00
method.call(globalObject, promise);
} else {
2016-01-09 02:07:10 +00:00
method.call(globalObject, reason, promise);
}
return true;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
})();
2015-06-08 06:38:56 +00:00
if (typeof console !== "undefined" && typeof console.warn !== "undefined") {
2016-01-09 02:07:10 +00:00
printWarning = function (message) {
2015-06-08 06:38:56 +00:00
console.warn(message);
};
if (util.isNode && process.stderr.isTTY) {
2016-01-09 02:07:10 +00:00
printWarning = function(message, isSoft) {
var color = isSoft ? "\u001b[33m" : "\u001b[31m";
console.warn(color + message + "\u001b[0m\n");
2015-06-08 06:38:56 +00:00
};
} else if (!util.isNode && typeof (new Error().stack) === "string") {
2016-01-09 02:07:10 +00:00
printWarning = function(message, isSoft) {
console.warn("%c" + message,
isSoft ? "color: darkorange" : "color: red");
2015-06-08 06:38:56 +00:00
};
}
}
2016-01-09 02:07:10 +00:00
var config = {
warnings: warnings,
longStackTraces: false,
cancellation: false
};
if (longStackTraces) Promise.longStackTraces();
return {
longStackTraces: function() {
return config.longStackTraces;
},
warnings: function() {
return config.warnings;
},
cancellation: function() {
return config.cancellation;
},
propagateFromFunction: function() {
return propagateFromFunction;
},
boundValueFunction: function() {
return boundValueFunction;
},
checkForgottenReturns: checkForgottenReturns,
setBounds: setBounds,
warn: warn,
deprecated: deprecated,
CapturedTrace: CapturedTrace
};
};
},{"./errors":12,"./util":36}],10:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
2016-01-09 02:07:10 +00:00
module.exports = function(Promise) {
function returner() {
return this.value;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
function thrower() {
throw this.reason;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
Promise.prototype["return"] =
Promise.prototype.thenReturn = function (value) {
if (value instanceof Promise) value.suppressUnhandledRejections();
return this._then(
returner, undefined, undefined, {value: value}, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype["throw"] =
Promise.prototype.thenThrow = function (reason) {
return this._then(
thrower, undefined, undefined, {reason: reason}, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype.catchThrow = function (reason) {
if (arguments.length <= 1) {
return this._then(
undefined, thrower, undefined, {reason: reason}, undefined);
} else {
var _reason = arguments[1];
var handler = function() {throw _reason;};
return this.caught(reason, handler);
}
};
2016-01-09 02:07:10 +00:00
Promise.prototype.catchReturn = function (value) {
if (arguments.length <= 1) {
if (value instanceof Promise) value.suppressUnhandledRejections();
return this._then(
undefined, returner, undefined, {value: value}, undefined);
} else {
var _value = arguments[1];
if (_value instanceof Promise) _value.suppressUnhandledRejections();
var handler = function() {return _value;};
return this.caught(value, handler);
}
};
};
2016-01-09 02:07:10 +00:00
},{}],11:[function(_dereq_,module,exports){
"use strict";
2016-01-09 02:07:10 +00:00
module.exports = function(Promise, INTERNAL) {
var PromiseReduce = Promise.reduce;
var PromiseAll = Promise.all;
2015-10-29 06:23:22 +00:00
2016-01-09 02:07:10 +00:00
function promiseAllThis() {
return PromiseAll(this);
2015-06-08 06:38:56 +00:00
}
2016-01-09 02:07:10 +00:00
function PromiseMapSeries(promises, fn) {
return PromiseReduce(promises, fn, INTERNAL, INTERNAL);
}
2016-01-09 02:07:10 +00:00
Promise.prototype.each = function (fn) {
return this.mapSeries(fn)
._then(promiseAllThis, undefined, undefined, this, undefined);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
Promise.prototype.mapSeries = function (fn) {
return PromiseReduce(this, fn, INTERNAL, INTERNAL);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
Promise.each = function (promises, fn) {
2016-01-09 02:07:10 +00:00
return PromiseMapSeries(promises, fn)
._then(promiseAllThis, undefined, undefined, promises, undefined);
};
2016-01-09 02:07:10 +00:00
Promise.mapSeries = PromiseMapSeries;
};
2016-01-09 02:07:10 +00:00
},{}],12:[function(_dereq_,module,exports){
"use strict";
2016-01-09 02:07:10 +00:00
var es5 = _dereq_("./es5");
2015-06-08 06:38:56 +00:00
var Objectfreeze = es5.freeze;
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
var inherits = util.inherits;
var notEnumerableProp = util.notEnumerableProp;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function subError(nameProperty, defaultMessage) {
function SubError(message) {
if (!(this instanceof SubError)) return new SubError(message);
notEnumerableProp(this, "message",
typeof message === "string" ? message : defaultMessage);
notEnumerableProp(this, "name", nameProperty);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (Error.captureStackTrace) {
Error.captureStackTrace(this, this.constructor);
2015-02-03 20:28:54 +00:00
} else {
Error.call(this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
inherits(SubError, Error);
return SubError;
}
var _TypeError, _RangeError;
2015-02-03 20:28:54 +00:00
var Warning = subError("Warning", "warning");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var CancellationError = subError("CancellationError", "cancellation error");
var TimeoutError = subError("TimeoutError", "timeout error");
var AggregateError = subError("AggregateError", "aggregate error");
try {
_TypeError = TypeError;
_RangeError = RangeError;
} catch(e) {
_TypeError = subError("TypeError", "type error");
_RangeError = subError("RangeError", "range error");
}
var methods = ("join pop push shift unshift slice filter forEach some " +
"every map indexOf lastIndexOf reduce reduceRight sort reverse").split(" ");
for (var i = 0; i < methods.length; ++i) {
if (typeof Array.prototype[methods[i]] === "function") {
AggregateError.prototype[methods[i]] = Array.prototype[methods[i]];
}
}
2015-06-08 06:38:56 +00:00
es5.defineProperty(AggregateError.prototype, "length", {
value: 0,
configurable: false,
writable: true,
enumerable: true
});
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
AggregateError.prototype["isOperational"] = true;
var level = 0;
AggregateError.prototype.toString = function() {
var indent = Array(level * 4 + 1).join(" ");
var ret = "\n" + indent + "AggregateError of:" + "\n";
level++;
indent = Array(level * 4 + 1).join(" ");
for (var i = 0; i < this.length; ++i) {
var str = this[i] === this ? "[Circular AggregateError]" : this[i] + "";
var lines = str.split("\n");
for (var j = 0; j < lines.length; ++j) {
lines[j] = indent + lines[j];
}
str = lines.join("\n");
ret += str + "\n";
}
level--;
return ret;
};
function OperationalError(message) {
if (!(this instanceof OperationalError))
return new OperationalError(message);
notEnumerableProp(this, "name", "OperationalError");
notEnumerableProp(this, "message", message);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this.cause = message;
this["isOperational"] = true;
if (message instanceof Error) {
notEnumerableProp(this, "message", message.message);
notEnumerableProp(this, "stack", message.stack);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else if (Error.captureStackTrace) {
Error.captureStackTrace(this, this.constructor);
}
}
inherits(OperationalError, Error);
var errorTypes = Error["__BluebirdErrorTypes__"];
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!errorTypes) {
errorTypes = Objectfreeze({
CancellationError: CancellationError,
TimeoutError: TimeoutError,
OperationalError: OperationalError,
RejectionError: OperationalError,
AggregateError: AggregateError
});
notEnumerableProp(Error, "__BluebirdErrorTypes__", errorTypes);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
module.exports = {
Error: Error,
TypeError: _TypeError,
RangeError: _RangeError,
CancellationError: errorTypes.CancellationError,
OperationalError: errorTypes.OperationalError,
TimeoutError: errorTypes.TimeoutError,
2015-02-03 20:28:54 +00:00
AggregateError: errorTypes.AggregateError,
Warning: Warning
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
},{"./es5":13,"./util":36}],13:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var isES5 = (function(){
"use strict";
return this === undefined;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
})();
if (isES5) {
module.exports = {
freeze: Object.freeze,
defineProperty: Object.defineProperty,
2015-06-08 06:38:56 +00:00
getDescriptor: Object.getOwnPropertyDescriptor,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
keys: Object.keys,
2015-06-08 06:38:56 +00:00
names: Object.getOwnPropertyNames,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
getPrototypeOf: Object.getPrototypeOf,
isArray: Array.isArray,
isES5: isES5,
propertyIsWritable: function(obj, prop) {
var descriptor = Object.getOwnPropertyDescriptor(obj, prop);
return !!(!descriptor || descriptor.writable || descriptor.set);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
} else {
var has = {}.hasOwnProperty;
var str = {}.toString;
var proto = {}.constructor.prototype;
var ObjectKeys = function (o) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = [];
for (var key in o) {
if (has.call(o, key)) {
ret.push(key);
}
}
return ret;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2015-06-08 06:38:56 +00:00
var ObjectGetDescriptor = function(o, key) {
return {value: o[key]};
};
var ObjectDefineProperty = function (o, key, desc) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
o[key] = desc.value;
return o;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ObjectFreeze = function (obj) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return obj;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ObjectGetPrototypeOf = function (obj) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
try {
return Object(obj).constructor.prototype;
}
catch (e) {
return proto;
}
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ArrayIsArray = function (obj) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
try {
return str.call(obj) === "[object Array]";
}
catch(e) {
return false;
}
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
module.exports = {
isArray: ArrayIsArray,
keys: ObjectKeys,
2015-06-08 06:38:56 +00:00
names: ObjectKeys,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
defineProperty: ObjectDefineProperty,
2015-06-08 06:38:56 +00:00
getDescriptor: ObjectGetDescriptor,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
freeze: ObjectFreeze,
getPrototypeOf: ObjectGetPrototypeOf,
isES5: isES5,
propertyIsWritable: function() {
return true;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
}
2016-01-09 02:07:10 +00:00
},{}],14:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise, INTERNAL) {
var PromiseMap = Promise.map;
Promise.prototype.filter = function (fn, options) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return PromiseMap(this, fn, options, INTERNAL);
};
Promise.filter = function (promises, fn, options) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return PromiseMap(promises, fn, options, INTERNAL);
};
};
2016-01-09 02:07:10 +00:00
},{}],15:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
2016-01-09 02:07:10 +00:00
module.exports = function(Promise, tryConvertToPromise) {
var util = _dereq_("./util");
var CancellationError = Promise.CancellationError;
var errorObj = util.errorObj;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
function FinallyHandlerCancelReaction(finallyHandler) {
this.finallyHandler = finallyHandler;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
FinallyHandlerCancelReaction.prototype._resultCancelled = function() {
checkCancel(this.finallyHandler);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
function checkCancel(ctx, reason) {
if (ctx.cancelPromise != null) {
if (arguments.length > 1) {
ctx.cancelPromise._reject(reason);
} else {
ctx.cancelPromise._cancel();
}
ctx.cancelPromise = null;
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
function succeed() {
return finallyHandler.call(this, this.promise._target()._settledValue());
}
function fail(reason) {
if (checkCancel(this, reason)) return;
errorObj.e = reason;
return errorObj;
}
function finallyHandler(reasonOrValue) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var promise = this.promise;
var handler = this.handler;
2016-01-09 02:07:10 +00:00
if (!this.called) {
this.called = true;
var ret = this.type === 0
? handler.call(promise._boundValue())
: handler.call(promise._boundValue(), reasonOrValue);
if (ret !== undefined) {
promise._setReturnedNonUndefined();
var maybePromise = tryConvertToPromise(ret, promise);
if (maybePromise instanceof Promise) {
if (this.cancelPromise != null) {
if (maybePromise.isCancelled()) {
var reason =
new CancellationError("late cancellation observer");
promise._attachExtraTrace(reason);
errorObj.e = reason;
return errorObj;
} else if (maybePromise.isPending()) {
maybePromise._attachCancellationCallback(
new FinallyHandlerCancelReaction(this));
}
}
return maybePromise._then(
succeed, fail, undefined, this, undefined);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
2016-01-09 02:07:10 +00:00
if (promise.isRejected()) {
checkCancel(this);
errorObj.e = reasonOrValue;
return errorObj;
} else {
checkCancel(this);
return reasonOrValue;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
Promise.prototype._passThrough = function(handler, type, success, fail) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof handler !== "function") return this.then();
2016-01-09 02:07:10 +00:00
return this._then(success, fail, undefined, {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
promise: this,
2016-01-09 02:07:10 +00:00
handler: handler,
called: false,
cancelPromise: null,
type: type
}, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.lastly =
Promise.prototype["finally"] = function (handler) {
2016-01-09 02:07:10 +00:00
return this._passThrough(handler,
0,
finallyHandler,
finallyHandler);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.tap = function (handler) {
2016-01-09 02:07:10 +00:00
return this._passThrough(handler, 1, finallyHandler);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
return finallyHandler;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],16:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise,
apiRejection,
INTERNAL,
2016-01-09 02:07:10 +00:00
tryConvertToPromise,
Proxyable,
debug) {
var errors = _dereq_("./errors");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var TypeError = errors.TypeError;
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var errorObj = util.errorObj;
var tryCatch = util.tryCatch;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var yieldHandlers = [];
function promiseFromYieldHandler(value, yieldHandlers, traceParent) {
for (var i = 0; i < yieldHandlers.length; ++i) {
traceParent._pushContext();
var result = tryCatch(yieldHandlers[i])(value);
traceParent._popContext();
if (result === errorObj) {
traceParent._pushContext();
var ret = Promise.reject(errorObj.e);
traceParent._popContext();
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
var maybePromise = tryConvertToPromise(result, traceParent);
if (maybePromise instanceof Promise) return maybePromise;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return null;
}
function PromiseSpawn(generatorFunction, receiver, yieldHandler, stack) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var promise = this._promise = new Promise(INTERNAL);
promise._captureStackTrace();
2016-01-09 02:07:10 +00:00
promise._setOnCancel(this);
this._stack = stack;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._generatorFunction = generatorFunction;
this._receiver = receiver;
this._generator = undefined;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._yieldHandlers = typeof yieldHandler === "function"
? [yieldHandler].concat(yieldHandlers)
: yieldHandlers;
2016-01-09 02:07:10 +00:00
this._yieldedPromise = null;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
util.inherits(PromiseSpawn, Proxyable);
PromiseSpawn.prototype._isResolved = function() {
return this._promise === null;
};
PromiseSpawn.prototype._cleanup = function() {
this._promise = this._generator = null;
};
PromiseSpawn.prototype._promiseCancelled = function() {
if (this._isResolved()) return;
var implementsReturn = typeof this._generator["return"] !== "undefined";
var result;
if (!implementsReturn) {
var reason = new Promise.CancellationError(
"generator .return() sentinel");
Promise.coroutine.returnSentinel = reason;
this._promise._attachExtraTrace(reason);
this._promise._pushContext();
result = tryCatch(this._generator["throw"]).call(this._generator,
reason);
this._promise._popContext();
if (result === errorObj && result.e === reason) {
result = null;
}
} else {
this._promise._pushContext();
result = tryCatch(this._generator["return"]).call(this._generator,
undefined);
this._promise._popContext();
}
var promise = this._promise;
this._cleanup();
if (result === errorObj) {
promise._rejectCallback(result.e, false);
} else {
promise.cancel();
}
};
PromiseSpawn.prototype._promiseFulfilled = function(value) {
this._yieldedPromise = null;
this._promise._pushContext();
var result = tryCatch(this._generator.next).call(this._generator, value);
this._promise._popContext();
this._continue(result);
};
PromiseSpawn.prototype._promiseRejected = function(reason) {
this._yieldedPromise = null;
this._promise._attachExtraTrace(reason);
this._promise._pushContext();
var result = tryCatch(this._generator["throw"])
.call(this._generator, reason);
this._promise._popContext();
this._continue(result);
};
PromiseSpawn.prototype._resultCancelled = function() {
if (this._yieldedPromise instanceof Promise) {
var promise = this._yieldedPromise;
this._yieldedPromise = null;
promise.cancel();
}
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
PromiseSpawn.prototype.promise = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._promise;
};
PromiseSpawn.prototype._run = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._generator = this._generatorFunction.call(this._receiver);
this._receiver =
this._generatorFunction = undefined;
2016-01-09 02:07:10 +00:00
this._promiseFulfilled(undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
PromiseSpawn.prototype._continue = function (result) {
2016-01-09 02:07:10 +00:00
var promise = this._promise;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (result === errorObj) {
2016-01-09 02:07:10 +00:00
this._cleanup();
return promise._rejectCallback(result.e, false);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
var value = result.value;
if (result.done === true) {
2016-01-09 02:07:10 +00:00
this._cleanup();
return promise._resolveCallback(value);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
var maybePromise = tryConvertToPromise(value, this._promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!(maybePromise instanceof Promise)) {
maybePromise =
promiseFromYieldHandler(maybePromise,
this._yieldHandlers,
this._promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (maybePromise === null) {
2016-01-09 02:07:10 +00:00
this._promiseRejected(
new TypeError(
2016-01-09 02:07:10 +00:00
"A value %s was yielded that could not be treated as a promise\u000a\u000a See http://goo.gl/MqrFmX\u000a\u000a".replace("%s", value) +
"From coroutine:\u000a" +
this._stack.split("\n").slice(1, -7).join("\n")
)
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
);
return;
}
}
2016-01-09 02:07:10 +00:00
maybePromise = maybePromise._target();
var bitField = maybePromise._bitField;
;
if (((bitField & 50397184) === 0)) {
this._yieldedPromise = maybePromise;
maybePromise._proxy(this, null);
} else if (((bitField & 33554432) !== 0)) {
this._promiseFulfilled(maybePromise._value());
} else if (((bitField & 16777216) !== 0)) {
this._promiseRejected(maybePromise._reason());
} else {
this._promiseCancelled();
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
Promise.coroutine = function (generatorFunction, options) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof generatorFunction !== "function") {
2016-01-09 02:07:10 +00:00
throw new TypeError("generatorFunction must be a function\u000a\u000a See http://goo.gl/MqrFmX\u000a");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
var yieldHandler = Object(options).yieldHandler;
var PromiseSpawn$ = PromiseSpawn;
var stack = new Error().stack;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return function () {
var generator = generatorFunction.apply(this, arguments);
var spawn = new PromiseSpawn$(undefined, undefined, yieldHandler,
stack);
2016-01-09 02:07:10 +00:00
var ret = spawn.promise();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
spawn._generator = generator;
2016-01-09 02:07:10 +00:00
spawn._promiseFulfilled(undefined);
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
Promise.coroutine.addYieldHandler = function(fn) {
2016-01-09 02:07:10 +00:00
if (typeof fn !== "function") {
throw new TypeError("expecting a function but got " + util.classString(fn));
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
yieldHandlers.push(fn);
};
Promise.spawn = function (generatorFunction) {
2016-01-09 02:07:10 +00:00
debug.deprecated("Promise.spawn()", "Promise.coroutine()");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof generatorFunction !== "function") {
2016-01-09 02:07:10 +00:00
return apiRejection("generatorFunction must be a function\u000a\u000a See http://goo.gl/MqrFmX\u000a");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
var spawn = new PromiseSpawn(generatorFunction, this);
var ret = spawn.promise();
spawn._run(Promise.spawn);
return ret;
};
};
2016-01-09 02:07:10 +00:00
},{"./errors":12,"./util":36}],17:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports =
function(Promise, PromiseArray, tryConvertToPromise, INTERNAL) {
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var canEvaluate = util.canEvaluate;
var tryCatch = util.tryCatch;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var errorObj = util.errorObj;
var reject;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!true) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (canEvaluate) {
var thenCallback = function(i) {
return new Function("value", "holder", " \n\
'use strict'; \n\
holder.pIndex = value; \n\
holder.checkFulfillment(this); \n\
".replace(/Index/g, i));
};
2016-01-09 02:07:10 +00:00
var promiseSetter = function(i) {
return new Function("promise", "holder", " \n\
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
'use strict'; \n\
2016-01-09 02:07:10 +00:00
holder.pIndex = promise; \n\
".replace(/Index/g, i));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
var generateHolderClass = function(total) {
var props = new Array(total);
for (var i = 0; i < props.length; ++i) {
props[i] = "this.p" + (i+1);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
var assignment = props.join(" = ") + " = null;";
var cancellationCode= "var promise;\n" + props.map(function(prop) {
return " \n\
promise = " + prop + "; \n\
if (promise instanceof Promise) { \n\
promise.cancel(); \n\
} \n\
";
}).join("\n");
var passedArguments = props.join(", ");
var name = "Holder$" + total;
var code = "return function(tryCatch, errorObj, Promise) { \n\
'use strict'; \n\
function [TheName](fn) { \n\
[TheProperties] \n\
this.fn = fn; \n\
this.now = 0; \n\
} \n\
[TheName].prototype.checkFulfillment = function(promise) { \n\
var now = ++this.now; \n\
if (now === [TheTotal]) { \n\
promise._pushContext(); \n\
var callback = this.fn; \n\
var ret = tryCatch(callback)([ThePassedArguments]); \n\
promise._popContext(); \n\
if (ret === errorObj) { \n\
promise._rejectCallback(ret.e, false); \n\
} else { \n\
promise._resolveCallback(ret); \n\
} \n\
} \n\
}; \n\
\n\
[TheName].prototype._resultCancelled = function() { \n\
[CancellationCode] \n\
}; \n\
\n\
return [TheName]; \n\
}(tryCatch, errorObj, Promise); \n\
";
code = code.replace(/\[TheName\]/g, name)
.replace(/\[TheTotal\]/g, total)
.replace(/\[ThePassedArguments\]/g, passedArguments)
.replace(/\[TheProperties\]/g, assignment)
.replace(/\[CancellationCode\]/g, cancellationCode);
return new Function("tryCatch", "errorObj", "Promise", code)
(tryCatch, errorObj, Promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
var holderClasses = [];
var thenCallbacks = [];
var promiseSetters = [];
for (var i = 0; i < 8; ++i) {
holderClasses.push(generateHolderClass(i + 1));
thenCallbacks.push(thenCallback(i + 1));
promiseSetters.push(promiseSetter(i + 1));
}
reject = function (reason) {
this._reject(reason);
};
2016-01-09 02:07:10 +00:00
}}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
Promise.join = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var last = arguments.length - 1;
var fn;
if (last > 0 && typeof arguments[last] === "function") {
fn = arguments[last];
if (!true) {
2016-01-09 02:07:10 +00:00
if (last <= 8 && canEvaluate) {
var ret = new Promise(INTERNAL);
ret._captureStackTrace();
2016-01-09 02:07:10 +00:00
var HolderClass = holderClasses[last - 1];
var holder = new HolderClass(fn);
var callbacks = thenCallbacks;
2016-01-09 02:07:10 +00:00
for (var i = 0; i < last; ++i) {
var maybePromise = tryConvertToPromise(arguments[i], ret);
if (maybePromise instanceof Promise) {
maybePromise = maybePromise._target();
2016-01-09 02:07:10 +00:00
var bitField = maybePromise._bitField;
;
if (((bitField & 50397184) === 0)) {
maybePromise._then(callbacks[i], reject,
undefined, ret, holder);
2016-01-09 02:07:10 +00:00
promiseSetters[i](maybePromise, holder);
} else if (((bitField & 33554432) !== 0)) {
callbacks[i].call(ret,
maybePromise._value(), holder);
2016-01-09 02:07:10 +00:00
} else if (((bitField & 16777216) !== 0)) {
ret._reject(maybePromise._reason());
2016-01-09 02:07:10 +00:00
} else {
ret._cancel();
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
callbacks[i].call(ret, maybePromise, holder);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
2016-01-09 02:07:10 +00:00
if (!ret._isFateSealed()) {
ret._setAsyncGuaranteed();
ret._setOnCancel(holder);
}
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
}
2016-01-09 02:07:10 +00:00
var args = [].slice.call(arguments);;
if (fn) args.pop();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = new PromiseArray(args).promise();
return fn !== undefined ? ret.spread(fn) : ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],18:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise,
PromiseArray,
apiRejection,
tryConvertToPromise,
2016-01-09 02:07:10 +00:00
INTERNAL,
debug) {
2015-10-29 06:23:22 +00:00
var getDomain = Promise._getDomain;
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
var tryCatch = util.tryCatch;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var errorObj = util.errorObj;
var EMPTY_ARRAY = [];
function MappingPromiseArray(promises, fn, limit, _filter) {
this.constructor$(promises);
this._promise._captureStackTrace();
2015-10-29 06:23:22 +00:00
var domain = getDomain();
this._callback = domain === null ? fn : domain.bind(fn);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._preservedValues = _filter === INTERNAL
? new Array(this.length())
: null;
this._limit = limit;
this._inFlight = 0;
this._queue = limit >= 1 ? [] : EMPTY_ARRAY;
2016-01-09 02:07:10 +00:00
this._init$(undefined, -2);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
util.inherits(MappingPromiseArray, PromiseArray);
MappingPromiseArray.prototype._init = function () {};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
MappingPromiseArray.prototype._promiseFulfilled = function (value, index) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var values = this._values;
var length = this.length();
var preservedValues = this._preservedValues;
var limit = this._limit;
2016-01-09 02:07:10 +00:00
if (index < 0) {
index = (index * -1) - 1;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
values[index] = value;
if (limit >= 1) {
this._inFlight--;
this._drainQueue();
2016-01-09 02:07:10 +00:00
if (this._isResolved()) return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
} else {
if (limit >= 1 && this._inFlight >= limit) {
values[index] = value;
this._queue.push(index);
2016-01-09 02:07:10 +00:00
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
if (preservedValues !== null) preservedValues[index] = value;
2016-01-09 02:07:10 +00:00
var promise = this._promise;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var callback = this._callback;
2016-01-09 02:07:10 +00:00
var receiver = promise._boundValue();
promise._pushContext();
var ret = tryCatch(callback).call(receiver, value, index, length);
2016-01-09 02:07:10 +00:00
var promiseCreated = promise._popContext();
debug.checkForgottenReturns(
ret,
promiseCreated,
preservedValues !== null ? "Promise.filter" : "Promise.map",
promise
);
if (ret === errorObj) {
this._reject(ret.e);
return true;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var maybePromise = tryConvertToPromise(ret, this._promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (maybePromise instanceof Promise) {
maybePromise = maybePromise._target();
2016-01-09 02:07:10 +00:00
var bitField = maybePromise._bitField;
;
if (((bitField & 50397184) === 0)) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (limit >= 1) this._inFlight++;
2016-01-09 02:07:10 +00:00
values[index] = maybePromise;
maybePromise._proxy(this, (index + 1) * -1);
return false;
} else if (((bitField & 33554432) !== 0)) {
ret = maybePromise._value();
2016-01-09 02:07:10 +00:00
} else if (((bitField & 16777216) !== 0)) {
this._reject(maybePromise._reason());
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
2016-01-09 02:07:10 +00:00
this._cancel();
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
values[index] = ret;
}
var totalResolved = ++this._totalResolved;
if (totalResolved >= length) {
if (preservedValues !== null) {
this._filter(values, preservedValues);
} else {
this._resolve(values);
}
2016-01-09 02:07:10 +00:00
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
MappingPromiseArray.prototype._drainQueue = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var queue = this._queue;
var limit = this._limit;
var values = this._values;
while (queue.length > 0 && this._inFlight < limit) {
if (this._isResolved()) return;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var index = queue.pop();
this._promiseFulfilled(values[index], index);
}
};
MappingPromiseArray.prototype._filter = function (booleans, values) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var len = values.length;
var ret = new Array(len);
var j = 0;
for (var i = 0; i < len; ++i) {
if (booleans[i]) ret[j++] = values[i];
}
ret.length = j;
this._resolve(ret);
};
MappingPromiseArray.prototype.preservedValues = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._preservedValues;
};
function map(promises, fn, options, _filter) {
2016-01-09 02:07:10 +00:00
if (typeof fn !== "function") {
return apiRejection("expecting a function but got " + util.classString(fn));
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var limit = typeof options === "object" && options !== null
? options.concurrency
: 0;
limit = typeof limit === "number" &&
isFinite(limit) && limit >= 1 ? limit : 0;
2016-01-09 02:07:10 +00:00
return new MappingPromiseArray(promises, fn, limit, _filter).promise();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
Promise.prototype.map = function (fn, options) {
2016-01-09 02:07:10 +00:00
return map(this, fn, options, null);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.map = function (promises, fn, options, _filter) {
2016-01-09 02:07:10 +00:00
return map(promises, fn, options, _filter);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],19:[function(_dereq_,module,exports){
"use strict";
module.exports =
2016-01-09 02:07:10 +00:00
function(Promise, INTERNAL, tryConvertToPromise, apiRejection, debug) {
var util = _dereq_("./util");
var tryCatch = util.tryCatch;
Promise.method = function (fn) {
if (typeof fn !== "function") {
2016-01-09 02:07:10 +00:00
throw new Promise.TypeError("expecting a function but got " + util.classString(fn));
}
return function () {
var ret = new Promise(INTERNAL);
ret._captureStackTrace();
ret._pushContext();
var value = tryCatch(fn).apply(this, arguments);
2016-01-09 02:07:10 +00:00
var promiseCreated = ret._popContext();
debug.checkForgottenReturns(
value, promiseCreated, "Promise.method", ret);
ret._resolveFromSyncValue(value);
return ret;
};
};
2016-01-09 02:07:10 +00:00
Promise.attempt = Promise["try"] = function (fn) {
if (typeof fn !== "function") {
2016-01-09 02:07:10 +00:00
return apiRejection("expecting a function but got " + util.classString(fn));
}
var ret = new Promise(INTERNAL);
ret._captureStackTrace();
ret._pushContext();
2016-01-09 02:07:10 +00:00
var value;
if (arguments.length > 1) {
debug.deprecated("calling Promise.try with more than 1 argument");
var arg = arguments[1];
var ctx = arguments[2];
value = util.isArray(arg) ? tryCatch(fn).apply(ctx, arg)
: tryCatch(fn).call(ctx, arg);
} else {
value = tryCatch(fn)();
}
var promiseCreated = ret._popContext();
debug.checkForgottenReturns(
value, promiseCreated, "Promise.try", ret);
ret._resolveFromSyncValue(value);
return ret;
};
Promise.prototype._resolveFromSyncValue = function (value) {
if (value === util.errorObj) {
2016-01-09 02:07:10 +00:00
this._rejectCallback(value.e, false);
} else {
this._resolveCallback(value, true);
}
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],20:[function(_dereq_,module,exports){
"use strict";
var util = _dereq_("./util");
var maybeWrapAsError = util.maybeWrapAsError;
var errors = _dereq_("./errors");
var OperationalError = errors.OperationalError;
var es5 = _dereq_("./es5");
function isUntypedError(obj) {
return obj instanceof Error &&
es5.getPrototypeOf(obj) === Error.prototype;
}
var rErrorKey = /^(?:name|message|stack|cause)$/;
function wrapAsOperationalError(obj) {
var ret;
if (isUntypedError(obj)) {
ret = new OperationalError(obj);
ret.name = obj.name;
ret.message = obj.message;
ret.stack = obj.stack;
var keys = es5.keys(obj);
for (var i = 0; i < keys.length; ++i) {
var key = keys[i];
if (!rErrorKey.test(key)) {
ret[key] = obj[key];
}
}
return ret;
}
util.markAsOriginatingFromRejection(obj);
return obj;
}
function nodebackForPromise(promise, multiArgs) {
return function(err, value) {
if (promise === null) return;
if (err) {
var wrapped = wrapAsOperationalError(maybeWrapAsError(err));
promise._attachExtraTrace(wrapped);
promise._reject(wrapped);
} else if (!multiArgs) {
promise._fulfill(value);
} else {
var args = [].slice.call(arguments, 1);;
promise._fulfill(args);
}
promise = null;
};
}
module.exports = nodebackForPromise;
},{"./errors":12,"./es5":13,"./util":36}],21:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise) {
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
var async = Promise._async;
var tryCatch = util.tryCatch;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var errorObj = util.errorObj;
function spreadAdapter(val, nodeback) {
var promise = this;
if (!util.isArray(val)) return successAdapter.call(promise, val, nodeback);
2015-10-29 06:23:22 +00:00
var ret =
tryCatch(nodeback).apply(promise._boundValue(), [null].concat(val));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (ret === errorObj) {
async.throwLater(ret.e);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
function successAdapter(val, nodeback) {
var promise = this;
2015-10-29 06:23:22 +00:00
var receiver = promise._boundValue();
var ret = val === undefined
? tryCatch(nodeback).call(receiver, null)
: tryCatch(nodeback).call(receiver, null, val);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (ret === errorObj) {
async.throwLater(ret.e);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
function errorAdapter(reason, nodeback) {
var promise = this;
if (!reason) {
2016-01-09 02:07:10 +00:00
var newReason = new Error(reason + "");
newReason.cause = reason;
reason = newReason;
}
2015-10-29 06:23:22 +00:00
var ret = tryCatch(nodeback).call(promise._boundValue(), reason);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (ret === errorObj) {
async.throwLater(ret.e);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
2016-01-09 02:07:10 +00:00
Promise.prototype.asCallback = Promise.prototype.nodeify = function (nodeback,
options) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof nodeback == "function") {
var adapter = successAdapter;
if (options !== undefined && Object(options).spread) {
adapter = spreadAdapter;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
this._then(
adapter,
errorAdapter,
undefined,
this,
nodeback
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
);
}
return this;
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],22:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function() {
var makeSelfResolutionError = function () {
2016-01-09 02:07:10 +00:00
return new TypeError("circular promise resolution chain\u000a\u000a See http://goo.gl/MqrFmX\u000a");
};
2016-01-09 02:07:10 +00:00
var reflectHandler = function() {
return new Promise.PromiseInspection(this._target());
};
var apiRejection = function(msg) {
return Promise.reject(new TypeError(msg));
};
2016-01-09 02:07:10 +00:00
function Proxyable() {}
var UNDEFINED_BINDING = {};
var util = _dereq_("./util");
2015-10-29 06:23:22 +00:00
var getDomain;
if (util.isNode) {
getDomain = function() {
var ret = process.domain;
if (ret === undefined) ret = null;
return ret;
};
} else {
getDomain = function() {
return null;
};
}
util.notEnumerableProp(Promise, "_getDomain", getDomain);
2016-01-09 02:07:10 +00:00
var es5 = _dereq_("./es5");
var Async = _dereq_("./async");
var async = new Async();
es5.defineProperty(Promise, "_async", {value: async});
var errors = _dereq_("./errors");
var TypeError = Promise.TypeError = errors.TypeError;
Promise.RangeError = errors.RangeError;
2016-01-09 02:07:10 +00:00
var CancellationError = Promise.CancellationError = errors.CancellationError;
Promise.TimeoutError = errors.TimeoutError;
Promise.OperationalError = errors.OperationalError;
Promise.RejectionError = errors.OperationalError;
Promise.AggregateError = errors.AggregateError;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var INTERNAL = function(){};
var APPLY = {};
2016-01-09 02:07:10 +00:00
var NEXT_FILTER = {};
var tryConvertToPromise = _dereq_("./thenables")(Promise, INTERNAL);
var PromiseArray =
2016-01-09 02:07:10 +00:00
_dereq_("./promise_array")(Promise, INTERNAL,
tryConvertToPromise, apiRejection, Proxyable);
var Context = _dereq_("./context")(Promise);
/*jshint unused:false*/
2016-01-09 02:07:10 +00:00
var createContext = Context.create;
var debug = _dereq_("./debuggability")(Promise, Context);
var CapturedTrace = debug.CapturedTrace;
var finallyHandler = _dereq_("./finally")(Promise, tryConvertToPromise);
var catchFilter = _dereq_("./catch_filter")(NEXT_FILTER);
var nodebackForPromise = _dereq_("./nodeback");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var errorObj = util.errorObj;
var tryCatch = util.tryCatch;
2016-01-09 02:07:10 +00:00
function check(self, executor) {
if (typeof executor !== "function") {
throw new TypeError("expecting a function but got " + util.classString(executor));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
if (self.constructor !== Promise) {
throw new TypeError("the promise constructor cannot be invoked directly\u000a\u000a See http://goo.gl/MqrFmX\u000a");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
}
function Promise(executor) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._bitField = 0;
this._fulfillmentHandler0 = undefined;
this._rejectionHandler0 = undefined;
this._promise0 = undefined;
this._receiver0 = undefined;
2016-01-09 02:07:10 +00:00
if (executor !== INTERNAL) {
check(this, executor);
this._resolveFromExecutor(executor);
}
this._promiseCreated();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
Promise.prototype.toString = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return "[object Promise]";
};
Promise.prototype.caught = Promise.prototype["catch"] = function (fn) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var len = arguments.length;
if (len > 1) {
var catchInstances = new Array(len - 1),
j = 0, i;
for (i = 0; i < len - 1; ++i) {
var item = arguments[i];
2016-01-09 02:07:10 +00:00
if (util.isObject(item)) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
catchInstances[j++] = item;
} else {
2016-01-09 02:07:10 +00:00
return apiRejection("expecting an object but got " + util.classString(item));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
catchInstances.length = j;
fn = arguments[i];
2016-01-09 02:07:10 +00:00
return this.then(undefined, catchFilter(catchInstances, fn, this));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return this.then(undefined, fn);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.reflect = function () {
2016-01-09 02:07:10 +00:00
return this._then(reflectHandler,
reflectHandler, undefined, this, undefined);
};
2016-01-09 02:07:10 +00:00
Promise.prototype.then = function (didFulfill, didReject) {
if (debug.warnings() && arguments.length > 0 &&
2015-02-03 20:28:54 +00:00
typeof didFulfill !== "function" &&
typeof didReject !== "function") {
var msg = ".then() only accepts functions but was passed: " +
util.classString(didFulfill);
if (arguments.length > 1) {
msg += ", " + util.classString(didReject);
}
this._warn(msg);
}
2016-01-09 02:07:10 +00:00
return this._then(didFulfill, didReject, undefined, undefined, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype.done = function (didFulfill, didReject) {
var promise =
this._then(didFulfill, didReject, undefined, undefined, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
promise._setIsFinal();
};
2016-01-09 02:07:10 +00:00
Promise.prototype.spread = function (fn) {
if (typeof fn !== "function") {
return apiRejection("expecting a function but got " + util.classString(fn));
}
return this.all()._then(fn, undefined, undefined, APPLY, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.toJSON = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = {
isFulfilled: false,
isRejected: false,
fulfillmentValue: undefined,
rejectionReason: undefined
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
if (this.isFulfilled()) {
ret.fulfillmentValue = this.value();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
ret.isFulfilled = true;
} else if (this.isRejected()) {
ret.rejectionReason = this.reason();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
ret.isRejected = true;
}
return ret;
};
Promise.prototype.all = function () {
2016-01-09 02:07:10 +00:00
if (arguments.length > 0) {
this._warn(".all() was passed arguments but it does not take any");
}
2015-06-08 06:38:56 +00:00
return new PromiseArray(this).promise();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.error = function (fn) {
return this.caught(util.originatesFromRejection, fn);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.is = function (val) {
return val instanceof Promise;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.fromNode = Promise.fromCallback = function(fn) {
var ret = new Promise(INTERNAL);
2016-01-09 02:07:10 +00:00
var multiArgs = arguments.length > 1 ? !!Object(arguments[1]).multiArgs
: false;
var result = tryCatch(fn)(nodebackForPromise(ret, multiArgs));
if (result === errorObj) {
2016-01-09 02:07:10 +00:00
ret._rejectCallback(result.e, true);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
if (!ret._isFateSealed()) ret._setAsyncGuaranteed();
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.all = function (promises) {
2015-06-08 06:38:56 +00:00
return new PromiseArray(promises).promise();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.cast = function (obj) {
var ret = tryConvertToPromise(obj);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!(ret instanceof Promise)) {
ret = new Promise(INTERNAL);
2016-01-09 02:07:10 +00:00
ret._captureStackTrace();
ret._setFulfilled();
ret._rejectionHandler0 = obj;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return ret;
};
Promise.resolve = Promise.fulfilled = Promise.cast;
Promise.reject = Promise.rejected = function (reason) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = new Promise(INTERNAL);
ret._captureStackTrace();
ret._rejectCallback(reason, true);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return ret;
};
Promise.setScheduler = function(fn) {
2016-01-09 02:07:10 +00:00
if (typeof fn !== "function") {
throw new TypeError("expecting a function but got " + util.classString(fn));
}
var prev = async._schedule;
async._schedule = fn;
return prev;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._then = function (
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
didFulfill,
didReject,
2016-01-09 02:07:10 +00:00
_, receiver,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
internalData
) {
var haveInternalData = internalData !== undefined;
2016-01-09 02:07:10 +00:00
var promise = haveInternalData ? internalData : new Promise(INTERNAL);
var target = this._target();
var bitField = target._bitField;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!haveInternalData) {
2016-01-09 02:07:10 +00:00
promise._propagateFrom(this, 3);
promise._captureStackTrace();
if (receiver === undefined &&
((this._bitField & 2097152) !== 0)) {
if (!((bitField & 50397184) === 0)) {
receiver = this._boundValue();
} else {
receiver = target === this ? undefined : this._boundTo;
}
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
var domain = getDomain();
if (!((bitField & 50397184) === 0)) {
var handler, value, settler = target._settlePromiseCtx;
if (((bitField & 33554432) !== 0)) {
value = target._rejectionHandler0;
handler = didFulfill;
} else if (((bitField & 16777216) !== 0)) {
value = target._fulfillmentHandler0;
handler = didReject;
target._unsetRejectionIsUnhandled();
} else {
settler = target._settlePromiseLateCancellationObserver;
value = new CancellationError("late cancellation observer");
target._attachExtraTrace(value);
handler = didReject;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
async.invoke(settler, target, {
handler: domain === null ? handler
: (typeof handler === "function" && domain.bind(handler)),
promise: promise,
receiver: receiver,
value: value
});
} else {
target._addCallbacks(didFulfill, didReject, promise, receiver, domain);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return promise;
};
Promise.prototype._length = function () {
2016-01-09 02:07:10 +00:00
return this._bitField & 65535;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._isFateSealed = function () {
return (this._bitField & 117506048) !== 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._isFollowing = function () {
2016-01-09 02:07:10 +00:00
return (this._bitField & 67108864) === 67108864;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._setLength = function (len) {
2016-01-09 02:07:10 +00:00
this._bitField = (this._bitField & -65536) |
(len & 65535);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._setFulfilled = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._bitField = this._bitField | 33554432;
};
2016-01-09 02:07:10 +00:00
Promise.prototype._setRejected = function () {
this._bitField = this._bitField | 16777216;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._setFollowing = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._bitField = this._bitField | 67108864;
};
2016-01-09 02:07:10 +00:00
Promise.prototype._setIsFinal = function () {
this._bitField = this._bitField | 4194304;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._isFinal = function () {
return (this._bitField & 4194304) > 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._unsetCancelled = function() {
this._bitField = this._bitField & (~65536);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._setCancelled = function() {
this._bitField = this._bitField | 65536;
};
Promise.prototype._setAsyncGuaranteed = function() {
this._bitField = this._bitField | 134217728;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._receiverAt = function (index) {
2016-01-09 02:07:10 +00:00
var ret = index === 0 ? this._receiver0 : this[
index * 4 - 4 + 3];
2015-10-29 06:23:22 +00:00
if (ret === UNDEFINED_BINDING) {
return undefined;
} else if (ret === undefined && this._isBound()) {
return this._boundValue();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return ret;
};
Promise.prototype._promiseAt = function (index) {
2016-01-09 02:07:10 +00:00
return this[
index * 4 - 4 + 2];
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._fulfillmentHandlerAt = function (index) {
2016-01-09 02:07:10 +00:00
return this[
index * 4 - 4 + 0];
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._rejectionHandlerAt = function (index) {
2016-01-09 02:07:10 +00:00
return this[
index * 4 - 4 + 1];
};
2016-01-09 02:07:10 +00:00
Promise.prototype._boundValue = function() {};
Promise.prototype._migrateCallback0 = function (follower) {
var bitField = follower._bitField;
var fulfill = follower._fulfillmentHandler0;
var reject = follower._rejectionHandler0;
var promise = follower._promise0;
var receiver = follower._receiverAt(0);
if (receiver === undefined) receiver = UNDEFINED_BINDING;
this._addCallbacks(fulfill, reject, promise, receiver, null);
2015-10-29 06:23:22 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._migrateCallbackAt = function (follower, index) {
var fulfill = follower._fulfillmentHandlerAt(index);
var reject = follower._rejectionHandlerAt(index);
var promise = follower._promiseAt(index);
var receiver = follower._receiverAt(index);
2015-10-29 06:23:22 +00:00
if (receiver === undefined) receiver = UNDEFINED_BINDING;
2016-01-09 02:07:10 +00:00
this._addCallbacks(fulfill, reject, promise, receiver, null);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._addCallbacks = function (
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
fulfill,
reject,
promise,
2015-10-29 06:23:22 +00:00
receiver,
domain
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
) {
var index = this._length();
2016-01-09 02:07:10 +00:00
if (index >= 65535 - 4) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
index = 0;
this._setLength(0);
}
if (index === 0) {
this._promise0 = promise;
2016-01-09 02:07:10 +00:00
this._receiver0 = receiver;
if (typeof fulfill === "function") {
2015-10-29 06:23:22 +00:00
this._fulfillmentHandler0 =
domain === null ? fulfill : domain.bind(fulfill);
}
if (typeof reject === "function") {
this._rejectionHandler0 =
domain === null ? reject : domain.bind(reject);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
2016-01-09 02:07:10 +00:00
var base = index * 4 - 4;
this[base + 2] = promise;
this[base + 3] = receiver;
2015-10-29 06:23:22 +00:00
if (typeof fulfill === "function") {
this[base + 0] =
domain === null ? fulfill : domain.bind(fulfill);
}
if (typeof reject === "function") {
this[base + 1] =
domain === null ? reject : domain.bind(reject);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
this._setLength(index + 1);
return index;
};
2016-01-09 02:07:10 +00:00
Promise.prototype._proxy = function (proxyable, arg) {
this._addCallbacks(undefined, undefined, arg, proxyable, null);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
Promise.prototype._resolveCallback = function(value, shouldBind) {
2016-01-09 02:07:10 +00:00
if (((this._bitField & 117506048) !== 0)) return;
if (value === this)
2016-01-09 02:07:10 +00:00
return this._rejectCallback(makeSelfResolutionError(), false);
var maybePromise = tryConvertToPromise(value, this);
if (!(maybePromise instanceof Promise)) return this._fulfill(value);
2016-01-09 02:07:10 +00:00
if (shouldBind) this._propagateFrom(maybePromise, 2);
var promise = maybePromise._target();
2016-01-09 02:07:10 +00:00
var bitField = promise._bitField;
if (((bitField & 50397184) === 0)) {
var len = this._length();
2016-01-09 02:07:10 +00:00
if (len > 0) promise._migrateCallback0(this);
for (var i = 1; i < len; ++i) {
promise._migrateCallbackAt(this, i);
}
this._setFollowing();
this._setLength(0);
this._setFollowee(promise);
2016-01-09 02:07:10 +00:00
} else if (((bitField & 33554432) !== 0)) {
this._fulfill(promise._value());
} else if (((bitField & 16777216) !== 0)) {
this._reject(promise._reason());
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
2016-01-09 02:07:10 +00:00
var reason = new CancellationError("late cancellation observer");
promise._attachExtraTrace(reason);
this._reject(reason);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
Promise.prototype._rejectCallback =
2016-01-09 02:07:10 +00:00
function(reason, synchronous, ignoreNonErrorWarnings) {
var trace = util.ensureErrorObject(reason);
2015-06-08 06:38:56 +00:00
var hasStack = trace === reason;
2016-01-09 02:07:10 +00:00
if (!hasStack && !ignoreNonErrorWarnings && debug.warnings()) {
var message = "a promise was rejected with a non-error: " +
util.classString(reason);
this._warn(message, true);
}
this._attachExtraTrace(trace, synchronous ? hasStack : false);
2016-01-09 02:07:10 +00:00
this._reject(reason);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._resolveFromExecutor = function (executor) {
var promise = this;
this._captureStackTrace();
this._pushContext();
var synchronous = true;
2016-01-09 02:07:10 +00:00
var r = this._execute(executor, function(value) {
promise._resolveCallback(value);
}, function (reason) {
promise._rejectCallback(reason, synchronous);
});
synchronous = false;
this._popContext();
2016-01-09 02:07:10 +00:00
if (r !== undefined) {
promise._rejectCallback(r, true);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
Promise.prototype._settlePromiseFromHandler = function (
handler, receiver, value, promise
) {
2016-01-09 02:07:10 +00:00
var bitField = promise._bitField;
if (((bitField & 65536) !== 0)) return;
promise._pushContext();
var x;
2016-01-09 02:07:10 +00:00
if (receiver === APPLY) {
if (!value || typeof value.length !== "number") {
x = errorObj;
x.e = new TypeError("cannot .spread() a non-array: " +
util.classString(value));
} else {
x = tryCatch(handler).apply(this._boundValue(), value);
}
} else {
x = tryCatch(handler).call(receiver, value);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
var promiseCreated = promise._popContext();
bitField = promise._bitField;
if (((bitField & 65536) !== 0)) return;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
if (x === NEXT_FILTER) {
promise._reject(value);
} else if (x === errorObj || x === promise) {
var err = x === promise ? makeSelfResolutionError() : x.e;
2016-01-09 02:07:10 +00:00
promise._rejectCallback(err, false);
} else {
2016-01-09 02:07:10 +00:00
debug.checkForgottenReturns(x, promiseCreated, "", promise, this);
promise._resolveCallback(x);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
Promise.prototype._target = function() {
var ret = this;
while (ret._isFollowing()) ret = ret._followee();
return ret;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
Promise.prototype._followee = function() {
return this._rejectionHandler0;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
Promise.prototype._setFollowee = function(promise) {
this._rejectionHandler0 = promise;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._settlePromise = function(promise, handler, receiver, value) {
var isPromise = promise instanceof Promise;
2016-01-09 02:07:10 +00:00
var bitField = this._bitField;
var asyncGuaranteed = ((bitField & 134217728) !== 0);
if (((bitField & 65536) !== 0)) {
if (isPromise) promise._invokeInternalOnCancel();
if (handler === finallyHandler) {
receiver.cancelPromise = promise;
if (tryCatch(handler).call(receiver, value) === errorObj) {
promise._reject(errorObj.e);
}
} else if (handler === reflectHandler) {
promise._fulfill(reflectHandler.call(receiver));
} else if (receiver instanceof Proxyable) {
receiver._promiseCancelled(promise);
} else if (isPromise || promise instanceof PromiseArray) {
promise._cancel();
} else {
receiver.cancel();
}
} else if (typeof handler === "function") {
if (!isPromise) {
handler.call(receiver, value, promise);
} else {
2016-01-09 02:07:10 +00:00
if (asyncGuaranteed) promise._setAsyncGuaranteed();
this._settlePromiseFromHandler(handler, receiver, value, promise);
}
2016-01-09 02:07:10 +00:00
} else if (receiver instanceof Proxyable) {
if (!receiver._isResolved()) {
2016-01-09 02:07:10 +00:00
if (((bitField & 33554432) !== 0)) {
receiver._promiseFulfilled(value, promise);
2016-01-09 02:07:10 +00:00
} else {
receiver._promiseRejected(value, promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
} else if (isPromise) {
2016-01-09 02:07:10 +00:00
if (asyncGuaranteed) promise._setAsyncGuaranteed();
if (((bitField & 33554432) !== 0)) {
promise._fulfill(value);
} else {
2016-01-09 02:07:10 +00:00
promise._reject(value);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
};
2016-01-09 02:07:10 +00:00
Promise.prototype._settlePromiseLateCancellationObserver = function(ctx) {
var handler = ctx.handler;
var promise = ctx.promise;
var receiver = ctx.receiver;
var value = ctx.value;
if (typeof handler === "function") {
if (!(promise instanceof Promise)) {
handler.call(receiver, value, promise);
} else {
this._settlePromiseFromHandler(handler, receiver, value, promise);
}
2016-01-09 02:07:10 +00:00
} else if (promise instanceof Promise) {
promise._reject(value);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._settlePromiseCtx = function(ctx) {
this._settlePromise(ctx.promise, ctx.handler, ctx.receiver, ctx.value);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._settlePromise0 = function(handler, value, bitField) {
var promise = this._promise0;
var receiver = this._receiverAt(0);
this._promise0 = undefined;
this._receiver0 = undefined;
this._settlePromise(promise, handler, receiver, value);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._clearCallbackDataAtIndex = function(index) {
var base = index * 4 - 4;
this[base + 2] =
this[base + 3] =
this[base + 0] =
this[base + 1] = undefined;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._fulfill = function (value) {
var bitField = this._bitField;
if (((bitField & 117506048) >>> 16)) return;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (value === this) {
var err = makeSelfResolutionError();
this._attachExtraTrace(err);
2016-01-09 02:07:10 +00:00
return this._reject(err);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
this._setFulfilled();
2016-01-09 02:07:10 +00:00
this._rejectionHandler0 = value;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
if ((bitField & 65535) > 0) {
if (((bitField & 134217728) !== 0)) {
this._settlePromises();
} else {
async.settlePromises(this);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
2016-01-09 02:07:10 +00:00
Promise.prototype._reject = function (reason) {
var bitField = this._bitField;
if (((bitField & 117506048) >>> 16)) return;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._setRejected();
2016-01-09 02:07:10 +00:00
this._fulfillmentHandler0 = reason;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (this._isFinal()) {
2016-01-09 02:07:10 +00:00
return async.fatalError(reason, util.isNode);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
if ((bitField & 65535) > 0) {
if (((bitField & 134217728) !== 0)) {
this._settlePromises();
} else {
async.settlePromises(this);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
this._ensurePossibleRejectionHandled();
}
};
2016-01-09 02:07:10 +00:00
Promise.prototype._fulfillPromises = function (len, value) {
for (var i = 1; i < len; i++) {
var handler = this._fulfillmentHandlerAt(i);
var promise = this._promiseAt(i);
var receiver = this._receiverAt(i);
this._clearCallbackDataAtIndex(i);
this._settlePromise(promise, handler, receiver, value);
}
};
Promise.prototype._rejectPromises = function (len, reason) {
for (var i = 1; i < len; i++) {
var handler = this._rejectionHandlerAt(i);
var promise = this._promiseAt(i);
var receiver = this._receiverAt(i);
this._clearCallbackDataAtIndex(i);
this._settlePromise(promise, handler, receiver, reason);
}
};
Promise.prototype._settlePromises = function () {
2016-01-09 02:07:10 +00:00
var bitField = this._bitField;
var len = (bitField & 65535);
if (len > 0) {
if (((bitField & 16842752) !== 0)) {
var reason = this._fulfillmentHandler0;
this._settlePromise0(this._rejectionHandler0, reason, bitField);
this._rejectPromises(len, reason);
} else {
var value = this._rejectionHandler0;
this._settlePromise0(this._fulfillmentHandler0, value, bitField);
this._fulfillPromises(len, value);
}
this._setLength(0);
}
this._clearCancellationData();
};
Promise.prototype._settledValue = function() {
var bitField = this._bitField;
if (((bitField & 33554432) !== 0)) {
return this._rejectionHandler0;
} else if (((bitField & 16777216) !== 0)) {
return this._fulfillmentHandler0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
2016-01-09 02:07:10 +00:00
function deferResolve(v) {this.promise._resolveCallback(v);}
function deferReject(v) {this.promise._rejectCallback(v, false);}
Promise.defer = Promise.pending = function() {
debug.deprecated("Promise.defer", "new Promise");
var promise = new Promise(INTERNAL);
return {
promise: promise,
resolve: deferResolve,
reject: deferReject
};
};
2015-10-29 06:23:22 +00:00
util.notEnumerableProp(Promise,
"_makeSelfResolutionError",
makeSelfResolutionError);
2016-01-09 02:07:10 +00:00
_dereq_("./method")(Promise, INTERNAL, tryConvertToPromise, apiRejection,
debug);
_dereq_("./bind")(Promise, INTERNAL, tryConvertToPromise, debug);
_dereq_("./cancel")(Promise, PromiseArray, apiRejection, debug);
_dereq_("./direct_resolve")(Promise);
_dereq_("./synchronous_inspection")(Promise);
_dereq_("./join")(
Promise, PromiseArray, tryConvertToPromise, INTERNAL, debug);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
Promise.Promise = Promise;
2016-01-09 02:07:10 +00:00
_dereq_('./map.js')(Promise, PromiseArray, apiRejection, tryConvertToPromise, INTERNAL, debug);
_dereq_('./using.js')(Promise, apiRejection, tryConvertToPromise, createContext, INTERNAL, debug);
_dereq_('./timers.js')(Promise, INTERNAL);
_dereq_('./generators.js')(Promise, apiRejection, INTERNAL, tryConvertToPromise, Proxyable, debug);
_dereq_('./nodeify.js')(Promise);
2015-06-08 06:38:56 +00:00
_dereq_('./call_get.js')(Promise);
_dereq_('./props.js')(Promise, PromiseArray, tryConvertToPromise, apiRejection);
_dereq_('./race.js')(Promise, INTERNAL, tryConvertToPromise, apiRejection);
2016-01-09 02:07:10 +00:00
_dereq_('./reduce.js')(Promise, PromiseArray, apiRejection, tryConvertToPromise, INTERNAL, debug);
_dereq_('./settle.js')(Promise, PromiseArray, debug);
_dereq_('./some.js')(Promise, PromiseArray, apiRejection);
2015-06-08 06:38:56 +00:00
_dereq_('./promisify.js')(Promise, INTERNAL);
_dereq_('./any.js')(Promise);
_dereq_('./each.js')(Promise, INTERNAL);
_dereq_('./filter.js')(Promise, INTERNAL);
2015-06-08 06:38:56 +00:00
util.toFastProperties(Promise);
util.toFastProperties(Promise.prototype);
function fillTypes(value) {
var p = new Promise(INTERNAL);
p._fulfillmentHandler0 = value;
p._rejectionHandler0 = value;
p._promise0 = value;
p._receiver0 = value;
}
// Complete slack tracking, opt out of field-type tracking and
// stabilize map
fillTypes({a: 1});
fillTypes({b: 2});
fillTypes({c: 3});
fillTypes(1);
fillTypes(function(){});
fillTypes(undefined);
fillTypes(false);
fillTypes(new Promise(INTERNAL));
2016-01-09 02:07:10 +00:00
debug.setBounds(Async.firstLineError, util.lastLineError);
2015-06-08 06:38:56 +00:00
return Promise;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
},{"./any.js":1,"./async":2,"./bind":3,"./call_get.js":5,"./cancel":6,"./catch_filter":7,"./context":8,"./debuggability":9,"./direct_resolve":10,"./each.js":11,"./errors":12,"./es5":13,"./filter.js":14,"./finally":15,"./generators.js":16,"./join":17,"./map.js":18,"./method":19,"./nodeback":20,"./nodeify.js":21,"./promise_array":23,"./promisify.js":24,"./props.js":25,"./race.js":27,"./reduce.js":28,"./settle.js":30,"./some.js":31,"./synchronous_inspection":32,"./thenables":33,"./timers.js":34,"./using.js":35,"./util":36}],23:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise, INTERNAL, tryConvertToPromise,
2016-01-09 02:07:10 +00:00
apiRejection, Proxyable) {
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var isArray = util.isArray;
function toResolutionValue(val) {
switch(val) {
case -2: return [];
case -3: return {};
}
}
function PromiseArray(values) {
var promise = this._promise = new Promise(INTERNAL);
if (values instanceof Promise) {
2016-01-09 02:07:10 +00:00
promise._propagateFrom(values, 3);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
promise._setOnCancel(this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._values = values;
this._length = 0;
this._totalResolved = 0;
this._init(undefined, -2);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
util.inherits(PromiseArray, Proxyable);
PromiseArray.prototype.length = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._length;
};
PromiseArray.prototype.promise = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._promise;
};
PromiseArray.prototype._init = function init(_, resolveValueIfEmpty) {
var values = tryConvertToPromise(this._values, this._promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (values instanceof Promise) {
values = values._target();
2016-01-09 02:07:10 +00:00
var bitField = values._bitField;
;
this._values = values;
2016-01-09 02:07:10 +00:00
if (((bitField & 50397184) === 0)) {
this._promise._setAsyncGuaranteed();
return values._then(
init,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._reject,
undefined,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this,
resolveValueIfEmpty
);
2016-01-09 02:07:10 +00:00
} else if (((bitField & 33554432) !== 0)) {
values = values._value();
} else if (((bitField & 16777216) !== 0)) {
return this._reject(values._reason());
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
2016-01-09 02:07:10 +00:00
return this._cancel();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
}
values = util.asArray(values);
if (values === null) {
var err = apiRejection(
"expecting an array or an iterable object but got " + util.classString(values)).reason();
this._promise._rejectCallback(err, false);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return;
}
if (values.length === 0) {
if (resolveValueIfEmpty === -5) {
this._resolveEmptyArray();
}
else {
this._resolve(toResolutionValue(resolveValueIfEmpty));
}
return;
}
2016-01-09 02:07:10 +00:00
this._iterate(values);
};
PromiseArray.prototype._iterate = function(values) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var len = this.getActualLength(values.length);
this._length = len;
this._values = this.shouldCopyValues() ? new Array(len) : this._values;
2016-01-09 02:07:10 +00:00
var result = this._promise;
var isResolved = false;
var bitField = null;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
for (var i = 0; i < len; ++i) {
2016-01-09 02:07:10 +00:00
var maybePromise = tryConvertToPromise(values[i], result);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (maybePromise instanceof Promise) {
maybePromise = maybePromise._target();
2016-01-09 02:07:10 +00:00
bitField = maybePromise._bitField;
} else {
bitField = null;
}
if (isResolved) {
if (bitField !== null) {
maybePromise.suppressUnhandledRejections();
}
} else if (bitField !== null) {
if (((bitField & 50397184) === 0)) {
maybePromise._proxy(this, i);
this._values[i] = maybePromise;
} else if (((bitField & 33554432) !== 0)) {
isResolved = this._promiseFulfilled(maybePromise._value(), i);
} else if (((bitField & 16777216) !== 0)) {
isResolved = this._promiseRejected(maybePromise._reason(), i);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
2016-01-09 02:07:10 +00:00
isResolved = this._promiseCancelled(i);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
} else {
isResolved = this._promiseFulfilled(maybePromise, i);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
2016-01-09 02:07:10 +00:00
if (!isResolved) result._setAsyncGuaranteed();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
PromiseArray.prototype._isResolved = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._values === null;
};
PromiseArray.prototype._resolve = function (value) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._values = null;
this._promise._fulfill(value);
};
2016-01-09 02:07:10 +00:00
PromiseArray.prototype._cancel = function() {
if (this._isResolved() || !this._promise.isCancellable()) return;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._values = null;
2016-01-09 02:07:10 +00:00
this._promise._cancel();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
PromiseArray.prototype._reject = function (reason) {
this._values = null;
this._promise._rejectCallback(reason, false);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
PromiseArray.prototype._promiseFulfilled = function (value, index) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._values[index] = value;
var totalResolved = ++this._totalResolved;
if (totalResolved >= this._length) {
this._resolve(this._values);
2016-01-09 02:07:10 +00:00
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
PromiseArray.prototype._promiseCancelled = function() {
this._cancel();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return true;
};
2016-01-09 02:07:10 +00:00
PromiseArray.prototype._promiseRejected = function (reason) {
this._totalResolved++;
this._reject(reason);
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
PromiseArray.prototype._resultCancelled = function() {
if (this._isResolved()) return;
var values = this._values;
this._cancel();
if (values instanceof Promise) {
values.cancel();
} else {
for (var i = 0; i < values.length; ++i) {
if (values[i] instanceof Promise) {
values[i].cancel();
}
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
2016-01-09 02:07:10 +00:00
PromiseArray.prototype.shouldCopyValues = function () {
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
PromiseArray.prototype.getActualLength = function (len) {
return len;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
return PromiseArray;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],24:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise, INTERNAL) {
var THIS = {};
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
var nodebackForPromise = _dereq_("./nodeback");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var withAppended = util.withAppended;
var maybeWrapAsError = util.maybeWrapAsError;
var canEvaluate = util.canEvaluate;
var TypeError = _dereq_("./errors").TypeError;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var defaultSuffix = "Async";
2015-06-08 06:38:56 +00:00
var defaultPromisified = {__isPromisified__: true};
2015-10-29 06:23:22 +00:00
var noCopyProps = [
"arity", "length",
"name",
"arguments",
"caller",
"callee",
"prototype",
"__isPromisified__"
];
var noCopyPropsPattern = new RegExp("^(?:" + noCopyProps.join("|") + ")$");
var defaultFilter = function(name) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return util.isIdentifier(name) &&
name.charAt(0) !== "_" &&
2015-10-29 06:23:22 +00:00
name !== "constructor";
};
2015-06-08 06:38:56 +00:00
function propsFilter(key) {
return !noCopyPropsPattern.test(key);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function isPromisified(fn) {
try {
return fn.__isPromisified__ === true;
}
catch (e) {
return false;
}
}
function hasPromisified(obj, key, suffix) {
var val = util.getDataPropertyOrDefault(obj, key + suffix,
defaultPromisified);
return val ? isPromisified(val) : false;
}
function checkValid(ret, suffix, suffixRegexp) {
for (var i = 0; i < ret.length; i += 2) {
var key = ret[i];
if (suffixRegexp.test(key)) {
var keyWithoutAsyncSuffix = key.replace(suffixRegexp, "");
for (var j = 0; j < ret.length; j += 2) {
if (ret[j] === keyWithoutAsyncSuffix) {
2016-01-09 02:07:10 +00:00
throw new TypeError("Cannot promisify an API that has normal methods with '%s'-suffix\u000a\u000a See http://goo.gl/MqrFmX\u000a"
.replace("%s", suffix));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
}
}
}
function promisifiableMethods(obj, suffix, suffixRegexp, filter) {
var keys = util.inheritedDataKeys(obj);
var ret = [];
for (var i = 0; i < keys.length; ++i) {
var key = keys[i];
var value = obj[key];
var passesDefaultFilter = filter === defaultFilter
? true : defaultFilter(key, value, obj);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof value === "function" &&
!isPromisified(value) &&
!hasPromisified(obj, key, suffix) &&
filter(key, value, obj, passesDefaultFilter)) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
ret.push(key, value);
}
}
checkValid(ret, suffix, suffixRegexp);
return ret;
}
var escapeIdentRegex = function(str) {
return str.replace(/([$])/, "\\$");
};
var makeNodePromisifiedEval;
if (!true) {
var switchCaseArgumentOrder = function(likelyArgumentCount) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = [likelyArgumentCount];
2015-06-08 06:38:56 +00:00
var min = Math.max(0, likelyArgumentCount - 1 - 3);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
for(var i = likelyArgumentCount - 1; i >= min; --i) {
ret.push(i);
}
2015-06-08 06:38:56 +00:00
for(var i = likelyArgumentCount + 1; i <= 3; ++i) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
ret.push(i);
}
return ret;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var argumentSequence = function(argumentCount) {
2015-06-08 06:38:56 +00:00
return util.filledRange(argumentCount, "_arg", "");
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var parameterDeclaration = function(parameterCount) {
2015-06-08 06:38:56 +00:00
return util.filledRange(
Math.max(parameterCount, 3), "_arg", "");
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var parameterCount = function(fn) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof fn.length === "number") {
return Math.max(Math.min(fn.length, 1023 + 1), 0);
}
return 0;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
makeNodePromisifiedEval =
2016-01-09 02:07:10 +00:00
function(callback, receiver, originalName, fn, _, multiArgs) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var newParameterCount = Math.max(0, parameterCount(fn) - 1);
var argumentOrder = switchCaseArgumentOrder(newParameterCount);
2015-06-08 06:38:56 +00:00
var shouldProxyThis = typeof callback === "string" || receiver === THIS;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function generateCallForArgumentCount(count) {
var args = argumentSequence(count).join(", ");
var comma = count > 0 ? ", " : "";
var ret;
2015-06-08 06:38:56 +00:00
if (shouldProxyThis) {
ret = "ret = callback.call(this, {{args}}, nodeback); break;\n";
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
2015-06-08 06:38:56 +00:00
ret = receiver === undefined
? "ret = callback({{args}}, nodeback); break;\n"
: "ret = callback.call(receiver, {{args}}, nodeback); break;\n";
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return ret.replace("{{args}}", args).replace(", ", comma);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
function generateArgumentSwitchCase() {
var ret = "";
2015-06-08 06:38:56 +00:00
for (var i = 0; i < argumentOrder.length; ++i) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
ret += "case " + argumentOrder[i] +":" +
generateCallForArgumentCount(argumentOrder[i]);
}
ret += " \n\
default: \n\
var args = new Array(len + 1); \n\
var i = 0; \n\
for (var i = 0; i < len; ++i) { \n\
args[i] = arguments[i]; \n\
} \n\
2015-06-08 06:38:56 +00:00
args[i] = nodeback; \n\
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
[CodeForCall] \n\
break; \n\
2015-06-08 06:38:56 +00:00
".replace("[CodeForCall]", (shouldProxyThis
? "ret = callback.apply(this, args);\n"
: "ret = callback.apply(receiver, args);\n"));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return ret;
}
2015-06-08 06:38:56 +00:00
var getFunctionCode = typeof callback === "string"
? ("this != null ? this['"+callback+"'] : fn")
: "fn";
2016-01-09 02:07:10 +00:00
var body = "'use strict'; \n\
2015-06-08 06:38:56 +00:00
var ret = function (Parameters) { \n\
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
'use strict'; \n\
var len = arguments.length; \n\
var promise = new Promise(INTERNAL); \n\
promise._captureStackTrace(); \n\
2016-01-09 02:07:10 +00:00
var nodeback = nodebackForPromise(promise, " + multiArgs + "); \n\
2015-06-08 06:38:56 +00:00
var ret; \n\
var callback = tryCatch([GetFunctionCode]); \n\
switch(len) { \n\
[CodeForSwitchCase] \n\
} \n\
if (ret === errorObj) { \n\
promise._rejectCallback(maybeWrapAsError(ret.e), true, true);\n\
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} \n\
2016-01-09 02:07:10 +00:00
if (!promise._isFateSealed()) promise._setAsyncGuaranteed(); \n\
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return promise; \n\
}; \n\
2015-10-29 06:23:22 +00:00
notEnumerableProp(ret, '__isPromisified__', true); \n\
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return ret; \n\
2016-01-09 02:07:10 +00:00
".replace("[CodeForSwitchCase]", generateArgumentSwitchCase())
.replace("[GetFunctionCode]", getFunctionCode);
body = body.replace("Parameters", parameterDeclaration(newParameterCount));
return new Function("Promise",
"fn",
"receiver",
"withAppended",
"maybeWrapAsError",
"nodebackForPromise",
"tryCatch",
"errorObj",
"notEnumerableProp",
"INTERNAL",
body)(
Promise,
fn,
receiver,
withAppended,
maybeWrapAsError,
nodebackForPromise,
util.tryCatch,
util.errorObj,
util.notEnumerableProp,
INTERNAL);
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
function makeNodePromisifiedClosure(callback, receiver, _, fn, __, multiArgs) {
2015-06-08 06:38:56 +00:00
var defaultThis = (function() {return this;})();
var method = callback;
if (typeof method === "string") {
callback = fn;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function promisified() {
var _receiver = receiver;
if (receiver === THIS) _receiver = this;
var promise = new Promise(INTERNAL);
promise._captureStackTrace();
2015-06-08 06:38:56 +00:00
var cb = typeof method === "string" && this !== defaultThis
? this[method] : callback;
2016-01-09 02:07:10 +00:00
var fn = nodebackForPromise(promise, multiArgs);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
try {
2015-06-08 06:38:56 +00:00
cb.apply(_receiver, withAppended(arguments, fn));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} catch(e) {
2015-06-08 06:38:56 +00:00
promise._rejectCallback(maybeWrapAsError(e), true, true);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
if (!promise._isFateSealed()) promise._setAsyncGuaranteed();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return promise;
}
2015-10-29 06:23:22 +00:00
util.notEnumerableProp(promisified, "__isPromisified__", true);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return promisified;
}
var makeNodePromisified = canEvaluate
? makeNodePromisifiedEval
: makeNodePromisifiedClosure;
2016-01-09 02:07:10 +00:00
function promisifyAll(obj, suffix, filter, promisifier, multiArgs) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var suffixRegexp = new RegExp(escapeIdentRegex(suffix) + "$");
var methods =
promisifiableMethods(obj, suffix, suffixRegexp, filter);
for (var i = 0, len = methods.length; i < len; i+= 2) {
var key = methods[i];
var fn = methods[i+1];
var promisifiedKey = key + suffix;
2015-10-29 06:23:22 +00:00
if (promisifier === makeNodePromisified) {
obj[promisifiedKey] =
2016-01-09 02:07:10 +00:00
makeNodePromisified(key, THIS, key, fn, suffix, multiArgs);
2015-10-29 06:23:22 +00:00
} else {
var promisified = promisifier(fn, function() {
2016-01-09 02:07:10 +00:00
return makeNodePromisified(key, THIS, key,
fn, suffix, multiArgs);
2015-10-29 06:23:22 +00:00
});
util.notEnumerableProp(promisified, "__isPromisified__", true);
obj[promisifiedKey] = promisified;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
util.toFastProperties(obj);
return obj;
}
2016-01-09 02:07:10 +00:00
function promisify(callback, receiver, multiArgs) {
return makeNodePromisified(callback, receiver, undefined,
callback, null, multiArgs);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
Promise.promisify = function (fn, options) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof fn !== "function") {
2016-01-09 02:07:10 +00:00
throw new TypeError("expecting a function but got " + util.classString(fn));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
if (isPromisified(fn)) {
return fn;
}
2016-01-09 02:07:10 +00:00
options = Object(options);
var receiver = options.context === undefined ? THIS : options.context;
var multiArgs = !!options.multiArgs;
var ret = promisify(fn, receiver, multiArgs);
2015-06-08 06:38:56 +00:00
util.copyDescriptors(fn, ret, propsFilter);
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.promisifyAll = function (target, options) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof target !== "function" && typeof target !== "object") {
2016-01-09 02:07:10 +00:00
throw new TypeError("the target of promisifyAll must be an object or a function\u000a\u000a See http://goo.gl/MqrFmX\u000a");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
options = Object(options);
2016-01-09 02:07:10 +00:00
var multiArgs = !!options.multiArgs;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var suffix = options.suffix;
if (typeof suffix !== "string") suffix = defaultSuffix;
var filter = options.filter;
if (typeof filter !== "function") filter = defaultFilter;
var promisifier = options.promisifier;
if (typeof promisifier !== "function") promisifier = makeNodePromisified;
if (!util.isIdentifier(suffix)) {
2016-01-09 02:07:10 +00:00
throw new RangeError("suffix must be a valid identifier\u000a\u000a See http://goo.gl/MqrFmX\u000a");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2015-06-08 06:38:56 +00:00
var keys = util.inheritedDataKeys(target);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
for (var i = 0; i < keys.length; ++i) {
var value = target[keys[i]];
if (keys[i] !== "constructor" &&
util.isClass(value)) {
2016-01-09 02:07:10 +00:00
promisifyAll(value.prototype, suffix, filter, promisifier,
multiArgs);
promisifyAll(value, suffix, filter, promisifier, multiArgs);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
2016-01-09 02:07:10 +00:00
return promisifyAll(target, suffix, filter, promisifier, multiArgs);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
2016-01-09 02:07:10 +00:00
},{"./errors":12,"./nodeback":20,"./util":36}],25:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(
Promise, PromiseArray, tryConvertToPromise, apiRejection) {
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var isObject = util.isObject;
2016-01-09 02:07:10 +00:00
var es5 = _dereq_("./es5");
var Es6Map;
if (typeof Map === "function") Es6Map = Map;
var mapToEntries = (function() {
var index = 0;
var size = 0;
function extractEntry(value, key) {
this[index] = value;
this[index + size] = key;
index++;
}
return function mapToEntries(map) {
size = map.size;
index = 0;
var ret = new Array(map.size * 2);
map.forEach(extractEntry, ret);
return ret;
};
})();
var entriesToMap = function(entries) {
var ret = new Es6Map();
var length = entries.length / 2 | 0;
for (var i = 0; i < length; ++i) {
var key = entries[length + i];
var value = entries[i];
ret.set(key, value);
}
return ret;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function PropertiesPromiseArray(obj) {
2016-01-09 02:07:10 +00:00
var isMap = false;
var entries;
if (Es6Map !== undefined && obj instanceof Es6Map) {
entries = mapToEntries(obj);
isMap = true;
} else {
var keys = es5.keys(obj);
var len = keys.length;
entries = new Array(len * 2);
for (var i = 0; i < len; ++i) {
var key = keys[i];
entries[i] = obj[key];
entries[i + len] = key;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
this.constructor$(entries);
this._isMap = isMap;
this._init$(undefined, -3);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
util.inherits(PropertiesPromiseArray, PromiseArray);
2016-01-09 02:07:10 +00:00
PropertiesPromiseArray.prototype._init = function () {};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
PropertiesPromiseArray.prototype._promiseFulfilled = function (value, index) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._values[index] = value;
var totalResolved = ++this._totalResolved;
if (totalResolved >= this._length) {
2016-01-09 02:07:10 +00:00
var val;
if (this._isMap) {
val = entriesToMap(this._values);
} else {
val = {};
var keyOffset = this.length();
for (var i = 0, len = this.length(); i < len; ++i) {
val[this._values[i + keyOffset]] = this._values[i];
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
this._resolve(val);
2016-01-09 02:07:10 +00:00
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
PropertiesPromiseArray.prototype.shouldCopyValues = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return false;
};
PropertiesPromiseArray.prototype.getActualLength = function (len) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return len >> 1;
};
function props(promises) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret;
var castValue = tryConvertToPromise(promises);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!isObject(castValue)) {
2016-01-09 02:07:10 +00:00
return apiRejection("cannot await properties of a non-object\u000a\u000a See http://goo.gl/MqrFmX\u000a");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else if (castValue instanceof Promise) {
ret = castValue._then(
Promise.props, undefined, undefined, undefined, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
ret = new PropertiesPromiseArray(castValue).promise();
}
if (castValue instanceof Promise) {
2016-01-09 02:07:10 +00:00
ret._propagateFrom(castValue, 2);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return ret;
}
Promise.prototype.props = function () {
return props(this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.props = function (promises) {
return props(promises);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
2016-01-09 02:07:10 +00:00
},{"./es5":13,"./util":36}],26:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
function arrayMove(src, srcIndex, dst, dstIndex, len) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
for (var j = 0; j < len; ++j) {
dst[j + dstIndex] = src[j + srcIndex];
src[j + srcIndex] = void 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
function Queue(capacity) {
this._capacity = capacity;
this._length = 0;
this._front = 0;
}
Queue.prototype._willBeOverCapacity = function (size) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._capacity < size;
};
Queue.prototype._pushOne = function (arg) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var length = this.length();
this._checkCapacity(length + 1);
var i = (this._front + length) & (this._capacity - 1);
this[i] = arg;
this._length = length + 1;
};
Queue.prototype._unshiftOne = function(value) {
var capacity = this._capacity;
this._checkCapacity(this.length() + 1);
var front = this._front;
var i = (((( front - 1 ) &
( capacity - 1) ) ^ capacity ) - capacity );
this[i] = value;
this._front = i;
this._length = this.length() + 1;
};
Queue.prototype.unshift = function(fn, receiver, arg) {
this._unshiftOne(arg);
this._unshiftOne(receiver);
this._unshiftOne(fn);
};
Queue.prototype.push = function (fn, receiver, arg) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var length = this.length() + 3;
if (this._willBeOverCapacity(length)) {
this._pushOne(fn);
this._pushOne(receiver);
this._pushOne(arg);
return;
}
var j = this._front + length - 3;
this._checkCapacity(length);
var wrapMask = this._capacity - 1;
this[(j + 0) & wrapMask] = fn;
this[(j + 1) & wrapMask] = receiver;
this[(j + 2) & wrapMask] = arg;
this._length = length;
};
Queue.prototype.shift = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var front = this._front,
ret = this[front];
this[front] = undefined;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._front = (front + 1) & (this._capacity - 1);
this._length--;
return ret;
};
Queue.prototype.length = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._length;
};
Queue.prototype._checkCapacity = function (size) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (this._capacity < size) {
this._resizeTo(this._capacity << 1);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
Queue.prototype._resizeTo = function (capacity) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var oldCapacity = this._capacity;
this._capacity = capacity;
var front = this._front;
var length = this._length;
var moveItemsCount = (front + length) & (oldCapacity - 1);
arrayMove(this, 0, this, oldCapacity, moveItemsCount);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
module.exports = Queue;
2016-01-09 02:07:10 +00:00
},{}],27:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(
Promise, INTERNAL, tryConvertToPromise, apiRejection) {
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var raceLater = function (promise) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return promise.then(function(array) {
return race(array, promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
});
};
function race(promises, parent) {
var maybePromise = tryConvertToPromise(promises);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (maybePromise instanceof Promise) {
return raceLater(maybePromise);
2016-01-09 02:07:10 +00:00
} else {
promises = util.asArray(promises);
if (promises === null)
return apiRejection("expecting an array or an iterable object but got " + util.classString(promises));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
var ret = new Promise(INTERNAL);
if (parent !== undefined) {
2016-01-09 02:07:10 +00:00
ret._propagateFrom(parent, 3);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
var fulfill = ret._fulfill;
var reject = ret._reject;
for (var i = 0, len = promises.length; i < len; ++i) {
var val = promises[i];
if (val === undefined && !(i in promises)) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
continue;
}
Promise.cast(val)._then(fulfill, reject, undefined, ret, null);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return ret;
}
Promise.race = function (promises) {
return race(promises, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.race = function () {
return race(this, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],28:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise,
PromiseArray,
apiRejection,
tryConvertToPromise,
2016-01-09 02:07:10 +00:00
INTERNAL,
debug) {
2015-10-29 06:23:22 +00:00
var getDomain = Promise._getDomain;
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
var tryCatch = util.tryCatch;
2016-01-09 02:07:10 +00:00
function ReductionPromiseArray(promises, fn, initialValue, _each) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this.constructor$(promises);
2015-10-29 06:23:22 +00:00
var domain = getDomain();
2016-01-09 02:07:10 +00:00
this._fn = domain === null ? fn : domain.bind(fn);
if (initialValue !== undefined) {
initialValue = Promise.resolve(initialValue);
initialValue._attachCancellationCallback(this);
}
this._initialValue = initialValue;
this._currentCancellable = null;
this._eachValues = _each === INTERNAL ? [] : undefined;
this._promise._captureStackTrace();
2015-06-08 06:38:56 +00:00
this._init$(undefined, -5);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
util.inherits(ReductionPromiseArray, PromiseArray);
2016-01-09 02:07:10 +00:00
ReductionPromiseArray.prototype._gotAccum = function(accum) {
if (this._eachValues !== undefined && accum !== INTERNAL) {
this._eachValues.push(accum);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
};
2016-01-09 02:07:10 +00:00
ReductionPromiseArray.prototype._eachComplete = function(value) {
this._eachValues.push(value);
return this._eachValues;
};
ReductionPromiseArray.prototype._init = function() {};
ReductionPromiseArray.prototype._resolveEmptyArray = function() {
this._resolve(this._eachValues !== undefined ? this._eachValues
: this._initialValue);
};
ReductionPromiseArray.prototype.shouldCopyValues = function () {
return false;
};
ReductionPromiseArray.prototype._resolve = function(value) {
this._promise._resolveCallback(value);
this._values = null;
};
ReductionPromiseArray.prototype._resultCancelled = function(sender) {
if (sender === this._initialValue) return this._cancel();
if (this._isResolved()) return;
this._resultCancelled$();
if (this._currentCancellable instanceof Promise) {
this._currentCancellable.cancel();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
if (this._initialValue instanceof Promise) {
this._initialValue.cancel();
}
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
ReductionPromiseArray.prototype._iterate = function (values) {
this._values = values;
var value;
var i;
var length = values.length;
if (this._initialValue !== undefined) {
value = this._initialValue;
i = 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else {
2016-01-09 02:07:10 +00:00
value = Promise.resolve(values[0]);
i = 1;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
this._currentCancellable = value;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
2016-01-09 02:07:10 +00:00
if (!value.isRejected()) {
for (; i < length; ++i) {
var ctx = {
accum: null,
value: values[i],
index: i,
length: length,
array: this
};
value = value._then(gotAccum, undefined, undefined, ctx, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
2016-01-09 02:07:10 +00:00
if (this._eachValues !== undefined) {
value = value
._then(this._eachComplete, undefined, undefined, this, undefined);
}
value._then(completed, completed, undefined, value, this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.reduce = function (fn, initialValue) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return reduce(this, fn, initialValue, null);
};
Promise.reduce = function (promises, fn, initialValue, _each) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return reduce(promises, fn, initialValue, _each);
};
2016-01-09 02:07:10 +00:00
function completed(valueOrReason, array) {
if (this.isFulfilled()) {
array._resolve(valueOrReason);
} else {
array._reject(valueOrReason);
}
}
function reduce(promises, fn, initialValue, _each) {
if (typeof fn !== "function") {
return apiRejection("expecting a function but got " + util.classString(fn));
}
var array = new ReductionPromiseArray(promises, fn, initialValue, _each);
return array.promise();
}
function gotAccum(accum) {
this.accum = accum;
this.array._gotAccum(accum);
var value = tryConvertToPromise(this.value, this.array._promise);
if (value instanceof Promise) {
this.array._currentCancellable = value;
return value._then(gotValue, undefined, undefined, this, undefined);
} else {
return gotValue.call(this, value);
}
}
function gotValue(value) {
var array = this.array;
var promise = array._promise;
var fn = tryCatch(array._fn);
promise._pushContext();
var ret;
if (array._eachValues !== undefined) {
ret = fn.call(promise._boundValue(), value, this.index, this.length);
} else {
ret = fn.call(promise._boundValue(),
this.accum, value, this.index, this.length);
}
if (ret instanceof Promise) {
array._currentCancellable = ret;
}
var promiseCreated = promise._popContext();
debug.checkForgottenReturns(
ret,
promiseCreated,
array._eachValues !== undefined ? "Promise.each" : "Promise.reduce",
promise
);
return ret;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],29:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
2015-06-08 06:38:56 +00:00
var util = _dereq_("./util");
2016-01-09 02:07:10 +00:00
var schedule;
2015-06-08 06:38:56 +00:00
var noAsyncScheduler = function() {
2016-01-09 02:07:10 +00:00
throw new Error("No async scheduler available\u000a\u000a See http://goo.gl/MqrFmX\u000a");
2015-06-08 06:38:56 +00:00
};
if (util.isNode && typeof MutationObserver === "undefined") {
var GlobalSetImmediate = global.setImmediate;
var ProcessNextTick = process.nextTick;
schedule = util.isRecentNode
? function(fn) { GlobalSetImmediate.call(global, fn); }
: function(fn) { ProcessNextTick.call(process, fn); };
2015-10-29 06:23:22 +00:00
} else if ((typeof MutationObserver !== "undefined") &&
!(typeof window !== "undefined" &&
window.navigator &&
window.navigator.standalone)) {
2016-01-09 02:07:10 +00:00
schedule = (function() {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var div = document.createElement("div");
2016-01-09 02:07:10 +00:00
var opts = {attributes: true};
var toggleScheduled = false;
var div2 = document.createElement("div");
var o2 = new MutationObserver(function() {
div.classList.toggle("foo");
toggleScheduled = false;
});
o2.observe(div2, opts);
var scheduleToggle = function() {
if (toggleScheduled) return;
toggleScheduled = true;
div2.classList.toggle("foo");
};
return function schedule(fn) {
var o = new MutationObserver(function() {
o.disconnect();
fn();
});
o.observe(div, opts);
scheduleToggle();
};
})();
2015-06-08 06:38:56 +00:00
} else if (typeof setImmediate !== "undefined") {
schedule = function (fn) {
2015-06-08 06:38:56 +00:00
setImmediate(fn);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2015-06-08 06:38:56 +00:00
} else if (typeof setTimeout !== "undefined") {
schedule = function (fn) {
setTimeout(fn, 0);
};
2015-06-08 06:38:56 +00:00
} else {
schedule = noAsyncScheduler;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
module.exports = schedule;
2016-01-09 02:07:10 +00:00
},{"./util":36}],30:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports =
2016-01-09 02:07:10 +00:00
function(Promise, PromiseArray, debug) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var PromiseInspection = Promise.PromiseInspection;
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function SettledPromiseArray(values) {
this.constructor$(values);
}
util.inherits(SettledPromiseArray, PromiseArray);
SettledPromiseArray.prototype._promiseResolved = function (index, inspection) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._values[index] = inspection;
var totalResolved = ++this._totalResolved;
if (totalResolved >= this._length) {
this._resolve(this._values);
2016-01-09 02:07:10 +00:00
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
SettledPromiseArray.prototype._promiseFulfilled = function (value, index) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = new PromiseInspection();
2016-01-09 02:07:10 +00:00
ret._bitField = 33554432;
ret._settledValueField = value;
return this._promiseResolved(index, ret);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
SettledPromiseArray.prototype._promiseRejected = function (reason, index) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = new PromiseInspection();
2016-01-09 02:07:10 +00:00
ret._bitField = 16777216;
ret._settledValueField = reason;
return this._promiseResolved(index, ret);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.settle = function (promises) {
2016-01-09 02:07:10 +00:00
debug.deprecated(".settle()", ".reflect()");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return new SettledPromiseArray(promises).promise();
};
Promise.prototype.settle = function () {
2016-01-09 02:07:10 +00:00
return Promise.settle(this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],31:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports =
function(Promise, PromiseArray, apiRejection) {
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
var RangeError = _dereq_("./errors").RangeError;
var AggregateError = _dereq_("./errors").AggregateError;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var isArray = util.isArray;
2016-01-09 02:07:10 +00:00
var CANCELLATION = {};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function SomePromiseArray(values) {
this.constructor$(values);
this._howMany = 0;
this._unwrap = false;
this._initialized = false;
}
util.inherits(SomePromiseArray, PromiseArray);
SomePromiseArray.prototype._init = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!this._initialized) {
return;
}
if (this._howMany === 0) {
this._resolve([]);
return;
}
this._init$(undefined, -5);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var isArrayResolved = isArray(this._values);
if (!this._isResolved() &&
isArrayResolved &&
this._howMany > this._canPossiblyFulfill()) {
this._reject(this._getRangeError(this.length()));
}
};
SomePromiseArray.prototype.init = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._initialized = true;
this._init();
};
SomePromiseArray.prototype.setUnwrap = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._unwrap = true;
};
SomePromiseArray.prototype.howMany = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._howMany;
};
SomePromiseArray.prototype.setHowMany = function (count) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._howMany = count;
};
SomePromiseArray.prototype._promiseFulfilled = function (value) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._addFulfilled(value);
if (this._fulfilled() === this.howMany()) {
this._values.length = this.howMany();
if (this.howMany() === 1 && this._unwrap) {
this._resolve(this._values[0]);
} else {
this._resolve(this._values);
}
2016-01-09 02:07:10 +00:00
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
SomePromiseArray.prototype._promiseRejected = function (reason) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._addRejected(reason);
2016-01-09 02:07:10 +00:00
return this._checkOutcome();
};
SomePromiseArray.prototype._promiseCancelled = function () {
if (this._values instanceof Promise || this._values == null) {
return this._cancel();
}
this._addRejected(CANCELLATION);
return this._checkOutcome();
};
SomePromiseArray.prototype._checkOutcome = function() {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (this.howMany() > this._canPossiblyFulfill()) {
var e = new AggregateError();
for (var i = this.length(); i < this._values.length; ++i) {
2016-01-09 02:07:10 +00:00
if (this._values[i] !== CANCELLATION) {
e.push(this._values[i]);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
if (e.length > 0) {
this._reject(e);
} else {
this._cancel();
}
return true;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
return false;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
SomePromiseArray.prototype._fulfilled = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._totalResolved;
};
SomePromiseArray.prototype._rejected = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._values.length - this.length();
};
SomePromiseArray.prototype._addRejected = function (reason) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._values.push(reason);
};
SomePromiseArray.prototype._addFulfilled = function (value) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._values[this._totalResolved++] = value;
};
SomePromiseArray.prototype._canPossiblyFulfill = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this.length() - this._rejected();
};
SomePromiseArray.prototype._getRangeError = function (count) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var message = "Input array must contain at least " +
this._howMany + " items but contains only " + count + " items";
return new RangeError(message);
};
SomePromiseArray.prototype._resolveEmptyArray = function () {
this._reject(this._getRangeError(0));
};
function some(promises, howMany) {
if ((howMany | 0) !== howMany || howMany < 0) {
2016-01-09 02:07:10 +00:00
return apiRejection("expecting a positive integer\u000a\u000a See http://goo.gl/MqrFmX\u000a");
}
var ret = new SomePromiseArray(promises);
var promise = ret.promise();
ret.setHowMany(howMany);
ret.init();
return promise;
}
Promise.some = function (promises, howMany) {
return some(promises, howMany);
};
Promise.prototype.some = function (howMany) {
return some(this, howMany);
};
Promise._SomePromiseArray = SomePromiseArray;
};
2016-01-09 02:07:10 +00:00
},{"./errors":12,"./util":36}],32:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise) {
function PromiseInspection(promise) {
if (promise !== undefined) {
promise = promise._target();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._bitField = promise._bitField;
2016-01-09 02:07:10 +00:00
this._settledValueField = promise._isFateSealed()
? promise._settledValue() : undefined;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
else {
this._bitField = 0;
2016-01-09 02:07:10 +00:00
this._settledValueField = undefined;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
2016-01-09 02:07:10 +00:00
PromiseInspection.prototype._settledValue = function() {
return this._settledValueField;
};
var value = PromiseInspection.prototype.value = function () {
if (!this.isFulfilled()) {
2016-01-09 02:07:10 +00:00
throw new TypeError("cannot get fulfillment value of a non-fulfilled promise\u000a\u000a See http://goo.gl/MqrFmX\u000a");
}
2016-01-09 02:07:10 +00:00
return this._settledValue();
};
2016-01-09 02:07:10 +00:00
var reason = PromiseInspection.prototype.error =
PromiseInspection.prototype.reason = function () {
if (!this.isRejected()) {
2016-01-09 02:07:10 +00:00
throw new TypeError("cannot get rejection reason of a non-rejected promise\u000a\u000a See http://goo.gl/MqrFmX\u000a");
}
2016-01-09 02:07:10 +00:00
return this._settledValue();
};
2016-01-09 02:07:10 +00:00
var isFulfilled = PromiseInspection.prototype.isFulfilled = function() {
return (this._bitField & 33554432) !== 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
var isRejected = PromiseInspection.prototype.isRejected = function () {
return (this._bitField & 16777216) !== 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
var isPending = PromiseInspection.prototype.isPending = function () {
return (this._bitField & 50397184) === 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
var isResolved = PromiseInspection.prototype.isResolved = function () {
return (this._bitField & 50331648) !== 0;
};
2016-01-09 02:07:10 +00:00
PromiseInspection.prototype.isCancelled =
Promise.prototype._isCancelled = function() {
return (this._bitField & 65536) === 65536;
};
2016-01-09 02:07:10 +00:00
Promise.prototype.isCancelled = function() {
return this._target()._isCancelled();
};
2016-01-09 02:07:10 +00:00
Promise.prototype.isPending = function() {
return isPending.call(this._target());
};
2016-01-09 02:07:10 +00:00
Promise.prototype.isRejected = function() {
return isRejected.call(this._target());
};
2016-01-09 02:07:10 +00:00
Promise.prototype.isFulfilled = function() {
return isFulfilled.call(this._target());
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype.isResolved = function() {
return isResolved.call(this._target());
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.value = function() {
2016-01-09 02:07:10 +00:00
return value.call(this._target());
};
Promise.prototype.reason = function() {
var target = this._target();
target._unsetRejectionIsUnhandled();
2016-01-09 02:07:10 +00:00
return reason.call(target);
};
Promise.prototype._value = function() {
return this._settledValue();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
Promise.prototype._reason = function() {
this._unsetRejectionIsUnhandled();
return this._settledValue();
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
Promise.PromiseInspection = PromiseInspection;
};
2016-01-09 02:07:10 +00:00
},{}],33:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise, INTERNAL) {
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var errorObj = util.errorObj;
var isObject = util.isObject;
function tryConvertToPromise(obj, context) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (isObject(obj)) {
2016-01-09 02:07:10 +00:00
if (obj instanceof Promise) return obj;
var then = getThen(obj);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (then === errorObj) {
if (context) context._pushContext();
var ret = Promise.reject(then.e);
if (context) context._popContext();
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} else if (typeof then === "function") {
2016-01-09 02:07:10 +00:00
if (isAnyBluebirdPromise(obj)) {
var ret = new Promise(INTERNAL);
obj._then(
ret._fulfill,
ret._reject,
undefined,
ret,
null
);
return ret;
}
return doThenable(obj, then, context);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
return obj;
}
2016-01-09 02:07:10 +00:00
function doGetThen(obj) {
return obj.then;
}
2016-01-09 02:07:10 +00:00
function getThen(obj) {
try {
return doGetThen(obj);
} catch (e) {
errorObj.e = e;
return errorObj;
}
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var hasProp = {}.hasOwnProperty;
function isAnyBluebirdPromise(obj) {
return hasProp.call(obj, "_promise0");
}
function doThenable(x, then, context) {
var promise = new Promise(INTERNAL);
var ret = promise;
if (context) context._pushContext();
promise._captureStackTrace();
if (context) context._popContext();
var synchronous = true;
2016-01-09 02:07:10 +00:00
var result = util.tryCatch(then).call(x, resolve, reject);
synchronous = false;
2016-01-09 02:07:10 +00:00
if (promise && result === errorObj) {
promise._rejectCallback(result.e, true, true);
promise = null;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
function resolve(value) {
if (!promise) return;
2015-10-29 06:23:22 +00:00
promise._resolveCallback(value);
promise = null;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
function reject(reason) {
if (!promise) return;
promise._rejectCallback(reason, synchronous, true);
promise = null;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return tryConvertToPromise;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],34:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function(Promise, INTERNAL) {
2016-01-09 02:07:10 +00:00
var util = _dereq_("./util");
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var TimeoutError = Promise.TimeoutError;
2016-01-09 02:07:10 +00:00
var afterTimeout = function (promise, message, parent) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (!promise.isPending()) return;
2015-10-29 06:23:22 +00:00
var err;
2016-01-09 02:07:10 +00:00
if (typeof message !== "string") {
if (message instanceof Error) {
err = message;
} else {
err = new TimeoutError("operation timed out");
2015-10-29 06:23:22 +00:00
}
2016-01-09 02:07:10 +00:00
} else {
2015-10-29 06:23:22 +00:00
err = new TimeoutError(message);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
util.markAsOriginatingFromRejection(err);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
promise._attachExtraTrace(err);
2016-01-09 02:07:10 +00:00
promise._reject(err);
parent.cancel();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
var afterValue = function(value) { return delay(+this).thenReturn(value); };
2016-01-09 02:07:10 +00:00
var delay = Promise.delay = function (ms, value) {
var ret;
if (value !== undefined) {
ret = Promise.resolve(value)
._then(afterValue, null, null, ms, undefined);
} else {
ret = new Promise(INTERNAL);
setTimeout(function() { ret._fulfill(); }, +ms);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
2016-01-09 02:07:10 +00:00
ret._setAsyncGuaranteed();
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.delay = function (ms) {
2016-01-09 02:07:10 +00:00
return delay(ms, this);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
function successClear(value) {
2015-02-03 20:28:54 +00:00
var handle = this;
if (handle instanceof Number) handle = +handle;
clearTimeout(handle);
return value;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function failureClear(reason) {
2015-02-03 20:28:54 +00:00
var handle = this;
if (handle instanceof Number) handle = +handle;
clearTimeout(handle);
throw reason;
}
Promise.prototype.timeout = function (ms, message) {
ms = +ms;
2016-01-09 02:07:10 +00:00
var parent = this.then();
var ret = parent.then();
var handle = setTimeout(function timeoutTimeout() {
2016-01-09 02:07:10 +00:00
afterTimeout(ret, message, parent);
}, ms);
return ret._then(successClear, failureClear, undefined, handle, undefined);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
};
2016-01-09 02:07:10 +00:00
},{"./util":36}],35:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
module.exports = function (Promise, apiRejection, tryConvertToPromise,
2016-01-09 02:07:10 +00:00
createContext, INTERNAL, debug) {
var util = _dereq_("./util");
var TypeError = _dereq_("./errors").TypeError;
var inherits = _dereq_("./util").inherits;
var errorObj = util.errorObj;
var tryCatch = util.tryCatch;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function thrower(e) {
setTimeout(function(){throw e;}, 0);
}
function castPreservingDisposable(thenable) {
var maybePromise = tryConvertToPromise(thenable);
if (maybePromise !== thenable &&
typeof thenable._isDisposable === "function" &&
typeof thenable._getDisposer === "function" &&
thenable._isDisposable()) {
maybePromise._setDisposable(thenable._getDisposer());
}
return maybePromise;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function dispose(resources, inspection) {
var i = 0;
var len = resources.length;
2016-01-09 02:07:10 +00:00
var ret = new Promise(INTERNAL);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function iterator() {
2016-01-09 02:07:10 +00:00
if (i >= len) return ret._fulfill();
var maybePromise = castPreservingDisposable(resources[i++]);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (maybePromise instanceof Promise &&
maybePromise._isDisposable()) {
try {
maybePromise = tryConvertToPromise(
maybePromise._getDisposer().tryDispose(inspection),
resources.promise);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
} catch (e) {
return thrower(e);
}
if (maybePromise instanceof Promise) {
return maybePromise._then(iterator, thrower,
null, null, null);
}
}
iterator();
}
iterator();
2016-01-09 02:07:10 +00:00
return ret;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
function Disposer(data, promise, context) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._data = data;
this._promise = promise;
this._context = context;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
Disposer.prototype.data = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._data;
};
Disposer.prototype.promise = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._promise;
};
Disposer.prototype.resource = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (this.promise().isFulfilled()) {
return this.promise().value();
}
return null;
};
Disposer.prototype.tryDispose = function(inspection) {
var resource = this.resource();
var context = this._context;
if (context !== undefined) context._pushContext();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = resource !== null
? this.doDispose(resource, inspection) : null;
if (context !== undefined) context._popContext();
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._promise._unsetDisposable();
this._data = null;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return ret;
};
Disposer.isDisposer = function (d) {
return (d != null &&
typeof d.resource === "function" &&
typeof d.tryDispose === "function");
};
function FunctionDisposer(fn, promise, context) {
this.constructor$(fn, promise, context);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
inherits(FunctionDisposer, Disposer);
FunctionDisposer.prototype.doDispose = function (resource, inspection) {
var fn = this.data();
return fn.call(resource, resource, inspection);
};
function maybeUnwrapDisposer(value) {
if (Disposer.isDisposer(value)) {
this.resources[this.index]._setDisposable(value);
return value.promise();
}
return value;
}
2016-01-09 02:07:10 +00:00
function ResourceList(length) {
this.length = length;
this.promise = null;
this[length-1] = null;
}
ResourceList.prototype._resultCancelled = function() {
var len = this.length;
for (var i = 0; i < len; ++i) {
var item = this[i];
if (item instanceof Promise) {
item.cancel();
}
}
};
Promise.using = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var len = arguments.length;
if (len < 2) return apiRejection(
"you must pass at least 2 arguments to Promise.using");
var fn = arguments[len - 1];
2016-01-09 02:07:10 +00:00
if (typeof fn !== "function") {
return apiRejection("expecting a function but got " + util.classString(fn));
}
2015-10-29 06:23:22 +00:00
var input;
var spreadArgs = true;
if (len === 2 && Array.isArray(arguments[0])) {
input = arguments[0];
len = input.length;
spreadArgs = false;
} else {
input = arguments;
len--;
}
2016-01-09 02:07:10 +00:00
var resources = new ResourceList(len);
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
for (var i = 0; i < len; ++i) {
2015-10-29 06:23:22 +00:00
var resource = input[i];
if (Disposer.isDisposer(resource)) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var disposer = resource;
resource = resource.promise();
resource._setDisposable(disposer);
} else {
var maybePromise = tryConvertToPromise(resource);
if (maybePromise instanceof Promise) {
resource =
maybePromise._then(maybeUnwrapDisposer, null, null, {
resources: resources,
index: i
}, undefined);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
resources[i] = resource;
}
2016-01-09 02:07:10 +00:00
var reflectedResources = new Array(resources.length);
for (var i = 0; i < reflectedResources.length; ++i) {
reflectedResources[i] = Promise.resolve(resources[i]).reflect();
}
var resultPromise = Promise.all(reflectedResources)
.then(function(inspections) {
for (var i = 0; i < inspections.length; ++i) {
var inspection = inspections[i];
if (inspection.isRejected()) {
errorObj.e = inspection.error();
return errorObj;
} else if (!inspection.isFulfilled()) {
resultPromise.cancel();
return;
}
inspections[i] = inspection.value();
}
2016-01-09 02:07:10 +00:00
promise._pushContext();
fn = tryCatch(fn);
var ret = spreadArgs
? fn.apply(undefined, inspections) : fn(inspections);
var promiseCreated = promise._popContext();
debug.checkForgottenReturns(
ret, promiseCreated, "Promise.using", promise);
return ret;
2016-01-09 02:07:10 +00:00
});
var promise = resultPromise.lastly(function() {
var inspection = new Promise.PromiseInspection(resultPromise);
return dispose(resources, inspection);
});
resources.promise = promise;
2016-01-09 02:07:10 +00:00
promise._setOnCancel(resources);
return promise;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._setDisposable = function (disposer) {
2016-01-09 02:07:10 +00:00
this._bitField = this._bitField | 131072;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
this._disposer = disposer;
};
Promise.prototype._isDisposable = function () {
2016-01-09 02:07:10 +00:00
return (this._bitField & 131072) > 0;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype._getDisposer = function () {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return this._disposer;
};
Promise.prototype._unsetDisposable = function () {
2016-01-09 02:07:10 +00:00
this._bitField = this._bitField & (~131072);
this._disposer = undefined;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
Promise.prototype.disposer = function (fn) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (typeof fn === "function") {
return new FunctionDisposer(fn, this, createContext());
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
throw new TypeError();
};
};
2016-01-09 02:07:10 +00:00
},{"./errors":12,"./util":36}],36:[function(_dereq_,module,exports){
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
"use strict";
2016-01-09 02:07:10 +00:00
var es5 = _dereq_("./es5");
var canEvaluate = typeof navigator == "undefined";
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var errorObj = {e: {}};
var tryCatchTarget;
function tryCatcher() {
try {
2015-10-29 06:23:22 +00:00
var target = tryCatchTarget;
tryCatchTarget = null;
return target.apply(this, arguments);
} catch (e) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
errorObj.e = e;
return errorObj;
}
}
function tryCatch(fn) {
tryCatchTarget = fn;
return tryCatcher;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
var inherits = function(Child, Parent) {
var hasProp = {}.hasOwnProperty;
function T() {
this.constructor = Child;
this.constructor$ = Parent;
for (var propertyName in Parent.prototype) {
if (hasProp.call(Parent.prototype, propertyName) &&
propertyName.charAt(propertyName.length-1) !== "$"
) {
this[propertyName + "$"] = Parent.prototype[propertyName];
}
}
}
T.prototype = Parent.prototype;
Child.prototype = new T();
return Child.prototype;
};
function isPrimitive(val) {
return val == null || val === true || val === false ||
typeof val === "string" || typeof val === "number";
}
function isObject(value) {
2016-01-09 02:07:10 +00:00
return typeof value === "function" ||
typeof value === "object" && value !== null;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
function maybeWrapAsError(maybeError) {
if (!isPrimitive(maybeError)) return maybeError;
2015-06-08 06:38:56 +00:00
return new Error(safeToString(maybeError));
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
function withAppended(target, appendee) {
var len = target.length;
var ret = new Array(len + 1);
var i;
for (i = 0; i < len; ++i) {
ret[i] = target[i];
}
ret[i] = appendee;
return ret;
}
function getDataPropertyOrDefault(obj, key, defaultValue) {
if (es5.isES5) {
var desc = Object.getOwnPropertyDescriptor(obj, key);
2015-10-29 06:23:22 +00:00
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (desc != null) {
return desc.get == null && desc.set == null
? desc.value
: defaultValue;
}
} else {
return {}.hasOwnProperty.call(obj, key) ? obj[key] : undefined;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
}
function notEnumerableProp(obj, name, value) {
if (isPrimitive(obj)) return obj;
var descriptor = {
value: value,
configurable: true,
enumerable: false,
writable: true
};
es5.defineProperty(obj, name, descriptor);
return obj;
}
function thrower(r) {
throw r;
}
var inheritedDataKeys = (function() {
2015-10-29 06:23:22 +00:00
var excludedPrototypes = [
Array.prototype,
Object.prototype,
Function.prototype
];
var isExcludedProto = function(val) {
for (var i = 0; i < excludedPrototypes.length; ++i) {
if (excludedPrototypes[i] === val) {
return true;
}
}
return false;
};
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
if (es5.isES5) {
2015-06-08 06:38:56 +00:00
var getKeys = Object.getOwnPropertyNames;
return function(obj) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = [];
var visitedKeys = Object.create(null);
2015-10-29 06:23:22 +00:00
while (obj != null && !isExcludedProto(obj)) {
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var keys;
try {
keys = getKeys(obj);
} catch (e) {
return ret;
}
for (var i = 0; i < keys.length; ++i) {
var key = keys[i];
if (visitedKeys[key]) continue;
visitedKeys[key] = true;
var desc = Object.getOwnPropertyDescriptor(obj, key);
if (desc != null && desc.get == null && desc.set == null) {
ret.push(key);
}
}
obj = es5.getPrototypeOf(obj);
}
return ret;
};
} else {
2015-10-29 06:23:22 +00:00
var hasProp = {}.hasOwnProperty;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
return function(obj) {
2015-10-29 06:23:22 +00:00
if (isExcludedProto(obj)) return [];
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = [];
2015-10-29 06:23:22 +00:00
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
/*jshint forin:false */
2015-10-29 06:23:22 +00:00
enumeration: for (var key in obj) {
if (hasProp.call(obj, key)) {
ret.push(key);
} else {
for (var i = 0; i < excludedPrototypes.length; ++i) {
if (hasProp.call(excludedPrototypes[i], key)) {
continue enumeration;
}
}
ret.push(key);
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return ret;
};
}
})();
2015-10-29 06:23:22 +00:00
var thisAssignmentPattern = /this\s*\.\s*\S+\s*=/;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
function isClass(fn) {
try {
if (typeof fn === "function") {
2015-06-08 06:38:56 +00:00
var keys = es5.names(fn.prototype);
2015-10-29 06:23:22 +00:00
var hasMethods = es5.isES5 && keys.length > 1;
var hasMethodsOtherThanConstructor = keys.length > 0 &&
!(keys.length === 1 && keys[0] === "constructor");
var hasThisAssignmentAndStaticMethods =
thisAssignmentPattern.test(fn + "") && es5.names(fn).length > 0;
if (hasMethods || hasMethodsOtherThanConstructor ||
hasThisAssignmentAndStaticMethods) {
return true;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
}
return false;
} catch (e) {
return false;
}
}
function toFastProperties(obj) {
2015-06-08 06:38:56 +00:00
/*jshint -W027,-W055,-W031*/
2016-01-09 02:07:10 +00:00
function FakeConstructor() {}
FakeConstructor.prototype = obj;
2015-06-08 06:38:56 +00:00
var l = 8;
2016-01-09 02:07:10 +00:00
while (l--) new FakeConstructor();
2015-06-08 06:38:56 +00:00
return obj;
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
eval(obj);
}
var rident = /^[a-z$_][a-z$_0-9]*$/i;
function isIdentifier(str) {
return rident.test(str);
}
function filledRange(count, prefix, suffix) {
var ret = new Array(count);
for(var i = 0; i < count; ++i) {
ret[i] = prefix + i + suffix;
}
return ret;
}
function safeToString(obj) {
try {
return obj + "";
} catch (e) {
return "[no string representation]";
}
}
function markAsOriginatingFromRejection(e) {
try {
notEnumerableProp(e, "isOperational", true);
}
catch(ignore) {}
}
function originatesFromRejection(e) {
if (e == null) return false;
return ((e instanceof Error["__BluebirdErrorTypes__"].OperationalError) ||
e["isOperational"] === true);
}
function canAttachTrace(obj) {
2015-10-29 06:23:22 +00:00
// Added by Zotero
return obj && obj.stack && es5.propertyIsWritable(obj, "stack");
2016-01-09 02:07:10 +00:00
return obj instanceof Error && es5.propertyIsWritable(obj, "stack");
}
var ensureErrorObject = (function() {
if (!("stack" in new Error())) {
return function(value) {
if (canAttachTrace(value)) return value;
try {throw new Error(safeToString(value));}
catch(err) {return err;}
};
} else {
return function(value) {
if (canAttachTrace(value)) return value;
return new Error(safeToString(value));
};
}
})();
2015-02-03 20:28:54 +00:00
function classString(obj) {
return {}.toString.call(obj);
}
2015-06-08 06:38:56 +00:00
function copyDescriptors(from, to, filter) {
var keys = es5.names(from);
for (var i = 0; i < keys.length; ++i) {
var key = keys[i];
if (filter(key)) {
2015-10-29 06:23:22 +00:00
try {
es5.defineProperty(to, key, es5.getDescriptor(from, key));
} catch (ignore) {}
2015-06-08 06:38:56 +00:00
}
}
}
2016-01-09 02:07:10 +00:00
var asArray = function(v) {
if (es5.isArray(v)) {
return v;
}
return null;
};
if (typeof Symbol !== "undefined" && Symbol.iterator) {
var ArrayFrom = typeof Array.from === "function" ? function(v) {
return Array.from(v);
} : function(v) {
var ret = [];
var it = v[Symbol.iterator]();
var itResult;
while (!((itResult = it.next()).done)) {
ret.push(itResult.value);
}
return ret;
};
asArray = function(v) {
if (es5.isArray(v)) {
return v;
} else if (v != null && typeof v[Symbol.iterator] === "function") {
return ArrayFrom(v);
}
return null;
};
}
var isNode = typeof process !== "undefined" &&
classString(process).toLowerCase() === "[object process]";
function env(key, def) {
return isNode ? process.env[key] : def;
}
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
var ret = {
isClass: isClass,
isIdentifier: isIdentifier,
inheritedDataKeys: inheritedDataKeys,
getDataPropertyOrDefault: getDataPropertyOrDefault,
thrower: thrower,
isArray: es5.isArray,
2016-01-09 02:07:10 +00:00
asArray: asArray,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
notEnumerableProp: notEnumerableProp,
isPrimitive: isPrimitive,
isObject: isObject,
canEvaluate: canEvaluate,
errorObj: errorObj,
tryCatch: tryCatch,
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
inherits: inherits,
withAppended: withAppended,
maybeWrapAsError: maybeWrapAsError,
toFastProperties: toFastProperties,
filledRange: filledRange,
toString: safeToString,
canAttachTrace: canAttachTrace,
ensureErrorObject: ensureErrorObject,
originatesFromRejection: originatesFromRejection,
2015-02-03 20:28:54 +00:00
markAsOriginatingFromRejection: markAsOriginatingFromRejection,
classString: classString,
2015-06-08 06:38:56 +00:00
copyDescriptors: copyDescriptors,
hasDevTools: typeof chrome !== "undefined" && chrome &&
typeof chrome.loadTimes === "function",
2016-01-09 02:07:10 +00:00
isNode: isNode,
env: env
Async DB megacommit Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details. WARNING: This includes backwards-incompatible schema changes. An incomplete list of other changes: - Schema overhaul - Replace main tables with new versions with updated schema - Enable real foreign key support and remove previous triggers - Don't use NULLs for local libraryID, which broke the UNIQUE index preventing object key duplication. All code (Zotero and third-party) using NULL for the local library will need to be updated to use 0 instead (already done for Zotero code) - Add 'compatibility' DB version that can be incremented manually to break DB compatibility with previous versions. 'userdata' upgrades will no longer automatically break compatibility. - Demote creators and tags from first-class objects to item properties - New API syncing properties - 'synced'/'version' properties to data objects - 'etag' to groups - 'version' to libraries - Create Zotero.DataObject that other objects inherit from - Consolidate data object loading into Zotero.DataObjects - Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data) - Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey() - New function Zotero.serial(fn), to wrap an async function such that all calls are run serially - New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func) - Add tag selector loading message - Various API and name changes, since everything was breaking anyway Known broken things: - Syncing (will be completely rewritten for API syncing) - Translation architecture (needs promise-based rewrite) - Duplicates view - DB integrity check (from schema changes) - Dragging (may be difficult to fix) Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
2014-08-06 21:38:05 +00:00
};
2015-06-08 06:38:56 +00:00
ret.isRecentNode = ret.isNode && (function() {
var version = process.versions.node.split(".").map(Number);
return (version[0] === 0 && version[1] > 10) || (version[0] > 0);
})();
2015-10-29 06:23:22 +00:00
if (ret.isNode) ret.toFastProperties(process);
2015-06-08 06:38:56 +00:00
2015-10-29 06:23:22 +00:00
try {throw new Error(); } catch (e) {ret.lastLineError = e;}
module.exports = ret;
2015-06-08 06:38:56 +00:00
2016-01-09 02:07:10 +00:00
},{"./es5":13}]},{},[4])(4)
}); ;if (typeof window !== 'undefined' && window !== null) { window.P = window.Promise; } else if (typeof self !== 'undefined' && self !== null) { self.P = self.Promise; }