Rather than requiring translators to explicitly set a referrer, as
proposed in #772 and #1375, this simply sets it to the URL where the
save button was triggered. This fixes the Project Euclid example
in #772. It's possible it won't fix all cases, since the translator might
build the URL manually or via an intermediate page, but hopefully it
will fix the majority of cases.
I guess there's a possibility that this would break something that
currently works, but it's hard to imagine a site would block based on
the wrong referrer from the right site and not block on no referrer.
Unlike #1375, this doesn't bother with the referrer for native downloads
(e.g., snapshots or images). The former probably don't need it, and the
latter should probably be switched to use `saveURI()` anyway.
This might also fixzotero/translators#523 (SSRN) if the translator
allowed it.
Closes#1375
In Firefox 52, `FileUtils.getFile()` tries to create the path to the
file even if the path array is empty. (In later versions (maybe in 54,
which we're using on macOS) it only does this if more than one path
component is provided.) We only need the special directory anyway, so we
can just use `getDir()`.
Follow-up to 4bbae6e17
We're now using Zotero.HTTP.request(), which does its own checking for
security errors, so there's no need to do WebDAV-specific checks (though
we could consider checking for Zotero.HTTP.SecurityError and showing
more specific messages, since a self-signed certificate is more likely
in the case of WebDAV).
If a server returns an invalid HTTP response (e.g., Content-Encoding:
gzip with a plaintext body, a.k.a. NS_ERROR_INVALID_CONTENT_ENCODING)
but we can still parse a 4xx or 5xx HTTP response code, use that for the
XHR status, since it might be enough for what we need to do (e.g.,
verify a 404 from a WebDAV server). This fixes a current problem with
Box [1].
Also fix a "msg is not defined" error when the XHR status is 0 but the
SSL connection is fine and include the channel and response status on
the UnexpectedStatusException object (though both of these would now
only happen on an invalid 2xx response, when the XHR status would remain
as 0).
[1] https://forums.zotero.org/discussion/comment/301014/#Comment_301014
purgeDataObjects(), which runs at the beginning of the sync process,
uses transactions, and those can fail after the default wait timeout if
there's another active transaction. Instead, check explicitly for
another transaction and, if there is one, display a nice message and
wait for it to finish. This isn't foolproof, but it should reduce the
frequency of "operation timed out" sync errors. (Avoiding all long
transactions would be a better solution.)
If a transaction took over 30 seconds and another transaction timed out
waiting for it, the second transaction would reset the notifier queue,
but if the first transaction then tried to queue an event, it would fail
with this error and roll back. (It would be nice to figure out why
transactions are taking over 30 seconds, though.)
- Updates /saveItems and /saveSnapshot to take a sessionID
- Provides a list of editable collections in the current library
- Adds an /updateSession method that takes a sessionID and updates the
collection and tags of any items saved via that operation (and changes
the currently selected collection)
Cross-library changes are not yet supported
I'm not sure what this was for, but at least with an async test function
it seems to be causing spurious "the string 'x' was thrown, throw an
Error :)" messages that hide the real error.
This is a simplified version of the fix from #872. Unlike the proposal
in #36, this doesn't require all child items to be selected, since in a
search some children might be grayed out. If the child of an unselected
parent item is included, the drag isn't allowed.
Closes#36
