Previously, objects were first downloaded and saved to the sync cache,
which was then processed separately to create/update local objects. This
meant that a server bug could result in invalid data in the sync cache
that would never be processed. Now, objects are saved as they're
downloaded and only added to the sync cache after being successfully
saved. The keys of objects that fail are added to a queue, and those
objects are refetched and retried on a backoff schedule or when a new
client version is installed (in case of a client bug or a client with
outdated data model support).
An alternative would be to save to the sync cache first and evict
objects that fail and add them to the queue, but that requires more
complicated logic, and it probably makes more sense just to buffer a few
downloads ahead so that processing is never waiting for downloads to
finish.
- Hide notes, tags and related for feed items in itembox
- Add feed support for <enclosure> elements
- Add feed syncing methods for synced settings (additional work is
needed on the sync architecture to download synced settings from the
server)
- Change feed item clear policy to be less aggressive
- Adjust for deasyncification
- Disable translate-on-select
- Closeadomasven/zotero#7, Remove context menu items from feeds
Fixes the incompatible-version dialog and adds info on the version used
to upgrade the database, which may be helpful for troubleshooting.
Also fixes showing of the Zotero toolbar icons even in case of a startup
error.
Also:
- Remove last-sync-time mechanism for both WebDAV and ZFS, since it can
be determined by storage properties (mtime/md5) in data sync
- Add option to include synced storage properties in item toJSON()
instead of local file properties
- Set "Fake-Server-Match" header in setHTTPResponse() test support
function, which can be used for request count assertions -- see
resetRequestCount() and assertRequestCount() in webdavTest.js
- Allow string (e.g., 'to_download') instead of constant in
Zotero.Sync.Data.Local.setSyncState()
- Misc storage tweaks
- Different constructor parameters
- id property for logging
- fcall() -> start()
- add() to enqueue without starting
- runAll() to run down queue and return promises for all current tasks
- wait() to wait for all running tasks to finish
Also:
* _finalizeErase in Zotero.DataObject is now inheritable
* Call _initErase before starting a DB transaction
* removes Zotero.Libraries.add and Zotero.Libraries.remove (doesn't seem like this is used any more)
This will appear much less frequently, since non-conflicting field changes on
both sides can be resolved automatically, but genuine field conflicts still
require manual conflict resolution.
The merge pane is no longer editable, since the itembox code to do that is
async and can't run in a modal window, but it's not really necessary,
particularly with conflicts happening less frequently.
TODO:
- Remote item deletions
- File conflicts
- Maybe handle some edge cases where the conflicted items fail to save
There's a lot more to do, and this isn't ready for actual usage, but the
basic functionality is mostly in place and has decent test coverage. It
can successfully upgrade a library last used with classic syncing and
pull down changes via the API. Uploading mostly works but is currently
disabled for safety until it has better test coverage.
Downloaded JSON is first saved to a cache table, which is then used to
populate other tables and later for generating PATCH requests and
automatically resolving conflicts (since it shows what was changed
locally and what was changed remotely). Objects with unmet dependencies
or unknown fields are skipped for now but don't block the rest of the
sync.
Some of the bigger remaining to-dos:
- Tests for uploading
- Re-do the preferences to get an API key
- File sync integration
- Full-text syncing integration
- Manual conflict resolution (though this already includes much smarter
conflict handling that automatically resolves many conflicts)
Since modal windows (e.g., the Create Bib window and the Quick Copy site
editor window) can't use yield, style retrieval
(Zotero.Styles.getVisible()/getAll()) is now synchronous, depending on a
previous async Zotero.Styles.init(). The translator list is generated in
the prefs window and passed into the Quick Copy site editor, but it's
possible the translators API should be changed to make getTranslators()
synchronous with a prior init() as well.
- Simplified schema
- Tags are now added without reloading entire tag selector
- On my system, adding 400 tags to an item (separately, with the tag
selector updating each time) went from 59 seconds to 42. (Given that
it takes only 13 seconds with the tag selector closed, though,
there's clearly more work to be done.)
- Tag selector now uses HTML flexbox (in identical fashion, for now, but
with the possibility of fancier changes later, and with streamlined
logic thanks to the flexbox 'order' property)
- Various async fixes
- Tests
Restore prepopulated charset table, but this time with just the
encodings from the WHATWG Encoding Standard. Assigning a charset to
Zotero.Item::attachmentCharset runs the value through
Zotero.CharacterSets.toCanonical() automatically.
This migrates attachment charsets to the new canonical values, clearing any
that are unsupported.
Other legacy mappings could still be added back, as disussed in #760.
And add group.fromJSON(json, userID), which sets editable and
filesEditable properties based on the group JSON (libraryReading, role
lists, etc.) and the given user
Relations are now properties of collections and items rather than
first-class objects, stored in separate collectionRelations and
itemRelations tables with ids for subjects, with foreign keys to the
associated data objects.
Related items now use dc:relation relations rather than a separate table
(among other reasons, because API syncing won't necessarily sync both
items at the same time, so they can't be stored by id).
The UI assigns related-item relations bidirectionally, and checks for
related-item and linked-object relations are done unidirectionally by
default.
dc:isReplacedBy is now dc:replaces, so that the subject is an existing
object, and the predicate is now named
Zotero.Attachments.replacedItemPredicate.
Some additional work is still needed, notably around following
replaced-item relations, and migration needs to be tested more fully,
but this seems to mostly work.
The test runner now downloads and caches the PDF tools for the current
platform within the test data directory and only redownloads them when
out of date, and it updates the download URL so that the full-text code
pulls from the cache directory via a file:// URL.
The installPDFTools() support function now installs the files directly
instead of going through the prefs, and a new uninstallPDFTools()
function removes the tools. Since the presence of the PDF tools can
affect other tests, tests that need the tools should install them in a
before() and uninstall them in an after(), leaving most tests to run
without PDF indexing.
This also adds a callback to the waitForWindow() support function. If a
modal dialog is opened, it blocks the next promise handler from running,
so a callback has to be used to interact with and close the dialog
immediately.
If a view or other resources are destroyed while a promise is being
resolved, subsequent code can fail. This is generally harmless, but it
results in unnecessary errors being logged to the console.
To address this, promises can use a new function,
Zotero.Promise.check(), to test whether a value is truthy or 0 and
automatically throw a specific error that's ignored by the unhandled
rejection handler if not.
Example usage:
getAsync().tap(() => Zotero.Promise.check(this.win));
If this.win is cleaned up while getAsync() is being resolved, subsequent
lines won't be run, and nothing will be logged to the console.
The sync cache will have pristine copies of the existing versions of
local objects for better conflict resolution, but downloads will get
saved to the sync cache first before processing, so the cache needs to
be able to hold more than one version.
And use 'version' instead of 'itemVersion' for object version for items
Also add deferred foreign key checking to system.sql so that DROP TABLE
commands don't fail mid-transaction
Copying zotero:// and other non-HTTP links from a note itself was fixed
in #452, but copying such links from other sources still stripped the
hrefs. This removes the patch in #452 and just gets the HTML directly
from the clipboard. I'm not sure why TinyMCE doesn't try to do this by
default (it only tries for plaintext), so maybe there's a problem with
this approach, but it seems to work for me (strips bad HTML, etc.).
Fixes#697
Use stdout redirection scripts for pdfinfo and, on Windows, a script to
run pdftotext hidden, which together allow for all unmodified binaries
(including, probably, symlinked system ones, though I didn't test that).
On Windows, using a .vbs does cause a brief wait cursor. The stock
pdfinfo needs the redirection script anyway, so that's unavoidable, but
on the async branch I think we'll be able to switch to pdf.js for the
page count, at which point maybe I'll try to remember how I modified the
Windows binaries to be hidden and use a modified version of pdftotext to
avoid VBScript. (We use the stock pdftotext elsewhere already.)
- Use async DB and OS.File for bundled file updates
- Remove support for translator/style ZIP files -- the two options are
now non-unpacked XPIs with subfolders or unpacked source installations
- Now that we have async file access, don't store translator code in
database cache -- just store metadata so that it's available without
reading each translator file
- Change the (previously partially asyncified) Zotero.Styles/Translators
APIs a bit -- while the getAll/getVisible methods are asynchronous and
will wait for loading, the get() methods are synchronous and require
styles/translators to be initialized before they're called. Most
places that end up calling get() probably call getAll/getVisible first
and should therefore be async, but if there's any way to trigger a
get() first, that will need to be adjusted.
- Asyncify various other style/translator-related code
XPI support is untested, as is style/translator usage, so there are
almost certainly bugs. The latter depends on updated export format
support (#659), since toArray() no longer exists on this branch.
Addresses #529 and #520
There are hundreds more, but these are all the ones that generate warnings in
the console at startup. XPCOM/XBL ones don't seem to do so, so we can ignore
those for now (and hopefully not bother with them on 4.0). Instances in
translators do generate warnings.
Addresses #656
Note that this loses conflicting changes to translate_item.js from 849803473a,
so those will need to be reapplied if applicable. /cc @aurimasv, @mtd91429
I didn't really investigate this, but errors from the outside world seem
to not pass the "obj instanceof Error" test due to some weirdness that
Bluebird is doing. (When I took this out recently I thought we just were
throwing some custom errors that didn't properly extend Error, but that
isn't the issue.) This restores the earlier workaround.
Instead of limiting charsets to a fixed list, dynamically populate it
with any charset name of less than 50 ASCII characters. Previously,
unknown charsets were discarded.
Zotero.Item.prototype.attachmentCharset now always returns a charset
name. It can be set with either a name or a charsetID.
Also:
- Remove the unused 'originalPath' column in itemAttachments
...with fewer modifications, since one (better debug info on an invalid
yield value) was implemented in Bluebird and the other (detecting a
thrown error that doesn't properly extend Error) we should just fix in
our code.
Our modifications no longer include a custom yield handler to
automatically call all() on yielded arrays (which maintained Bluebird
1.x behavior). It's now necessary to call all() or similar explicitly.
Also fixed a few incorrect yields hidden by that behavior.
- Protocol handler extensions can now handle promises and can also make
data available as it's ready instead of all at once (e.g., reports now
output one entry at a time)
- zotero:// URL syntaxes are now more consistent and closer to the web
API (old URLs should work, but some may currently be broken)
Also:
- Code to generate server API, currently available for testing via
zotero://data URLs but eventually moving to HTTP -- zotero://data URLs match
web API URLs, with a different prefix for the personal library (/library vs.
/users/12345)
- Miscellaneous fixes to data objects
Under the hood:
- Extensions now return an AsyncChannel, which is an nsIChannel implementation
that takes a promise-yielding generator that returns a string,
nsIAsyncInputStream, or file that will be used for the channel's data
- New function Zotero.Utilities.Internal.getAsyncInputStream() takes a
generator that yields either promises or strings and returns an async input
stream filled with the yielded strings
- Zotero.Router parsers URLs and extract parameters
- Zotero.Item.toResponseJSON()
This required doing additional caching at startup (e.g., item types and fields)
so that various methods can remain synchronous.
This lets us switch back to using the current Sqlite.jsm. Previously we were
bundling the Fx24 version, which avoided freezes with locking_mode=EXCLUSIVE
with both sync and async queries.
Known broken things:
- Autocomplete
- Database backup
- UDFs (e.g., REGEXP function used in Zotero.DB.getNextName())
Promise-based rewrite of most of the codebase, with asynchronous database and file access -- see https://github.com/zotero/zotero/issues/518 for details.
WARNING: This includes backwards-incompatible schema changes.
An incomplete list of other changes:
- Schema overhaul
- Replace main tables with new versions with updated schema
- Enable real foreign key support and remove previous triggers
- Don't use NULLs for local libraryID, which broke the UNIQUE index
preventing object key duplication. All code (Zotero and third-party)
using NULL for the local library will need to be updated to use 0
instead (already done for Zotero code)
- Add 'compatibility' DB version that can be incremented manually to break DB
compatibility with previous versions. 'userdata' upgrades will no longer
automatically break compatibility.
- Demote creators and tags from first-class objects to item properties
- New API syncing properties
- 'synced'/'version' properties to data objects
- 'etag' to groups
- 'version' to libraries
- Create Zotero.DataObject that other objects inherit from
- Consolidate data object loading into Zotero.DataObjects
- Change object reloading so that only the loaded and changed parts of objects are reloaded, instead of reloading all data from the database (with some exceptions, including item primary data)
- Items and collections now have .parentItem and .parentKey properties, replacing item.getSource() and item.getSourceKey()
- New function Zotero.serial(fn), to wrap an async function such that all calls are run serially
- New function Zotero.Utilities.Internal.forEachChunkAsync(arr, chunkSize, func)
- Add tag selector loading message
- Various API and name changes, since everything was breaking anyway
Known broken things:
- Syncing (will be completely rewritten for API syncing)
- Translation architecture (needs promise-based rewrite)
- Duplicates view
- DB integrity check (from schema changes)
- Dragging (may be difficult to fix)
Lots of other big and little things are certainly broken, particularly with the UI, which can be affected by async code in all sorts of subtle ways.
With Task.spawn, regular expressions in Zotero.DB were causing "too much
recursion" errors on Windows with JIT enabled.
This requires a change to Q to allow async() to take a generator instead
of a generator-maker (which is the reason it was using Task.spawn to
begin with).
- Send modifier keys through to loadURI() when clicking Open Link in notes
- Open link in parent window from external note window
- Don't show both menus on right-click
Follow-up from #450
Full-text content is now synced to the server and other clients. Up to
~500K of content is transferred with each sync, so multiple syncs may be
necessary. Downloaded content is written to temporary files and
processed on idle (currently 5 seconds), so synced content may not be
immediately available.
Performance will improve with API syncing and async DB in 4.1.
This also includes some other full-text tweaks, such as storing
text converted from HTML in cache files, which may speed up phrase
searching.
Remove overrides for titles we now abbreviate correctly, and don't
delete potential prepositions that are rarely used as such in journal
titles ("hyphen", "midst", "minus", "modulo", outside", "pace", "past",
"plus", "round", "save", "til", "till", "times", "vice", "worth")
This change should improve Firefox startup time. If the Zotero pane is
opened before Zotero has initialized, the pane will be blocked with a
progress bar until it's done. Standalone currently does the same, but it
should be changed to just delay opening the window if possible.
The upgrade wizard has been disabled for schema upgrades, in the hope
that upgrades can be done in a way that won't break compatibility with
earlier versions (or that can at least be done in a way such that we can
put out point releases of the last major version that provide
compatibility during beta/post-upgrade periods).
This patch likely breaks many things, and definitely breaks connector
mode.
This change also removes upgrade steps for databases from Zotero 2.1b2
and earlier. Users with such databases will need to upgrade via Zotero
4.0.x first or delete their data directories and start anew. This should
only affect users who haven't opened Zotero since Nov. 2010.
Task.jsm allows not only generators but also iterators to be yielded,
using ``typeof yielded == 'Function'`` as the test for an iterator, but
since the Q promise includes a send() method, yielding a Q promise
instead of a Mozilla promise results in an infinite loop. To fix this,
we disable Q's send() method, which is just an outdated synonym for
invoke() anyway.
This will have to be done every time we update Q until send() is
removed, Task.jsm is changed, or we stop using Task.jsm with Q promises.
It appears that this flag can no longer be set in new-ish Nightlies.
This patch switches code that used nsITimers with the
Components.utils.methodjit hack to use setTimeout on the hidden DOM
window instead, so that we still get the JIT. (See
https://bugzilla.mozilla.org/show_bug.cgi?id=776798)
This might also mean that callbacks from doGet etc. no longer get JITed.
At some point, we should test this with the JIT profiler.
- New tag colors support, with the ability to assign colors to up to 6
tags per library. Tags with colors assigned will show up at the top of
the tag selector and can be added to (and removed from) selected items
by pressing the 1-6 keys on the keyboard. The tags will show up as
color swatches before an item's title in the items list.
- Synced settings, with Notifier triggers when they change and
accessible via the API (currently restricted on the server to
'tagColors', but available for other things upon request)
- Silent DB upgrades for backwards-compatible changes. We'll do
something fancier with async DB queries in 4.0, but this will work for
changes that can be made without breaking compatibility with older
clients, like the creation of new tables. The 'userdata' value is
capped at 76, while further increments go to 'userdata2'.
TODO:
- Try to avoid jitter when redrawing swatches
- Optimize tag color images for retina displays
- Redo attachment dots in flat style?
- Clear all colors from an item with 0 (as in Thunderbird), but I don't
think we can do this without undo
