relinkAttachment() calls getClosestDirectory() before showing the file
picker, and a too-long filename caused the OS.File.stat() in
getClosestDirectory() to throw, which prevented the file picker from
appearing after clicking Locate.
https://forums.zotero.org/discussion/133685/
Firefox ESR 140.15 rejects loads from a content process for URLs that
process couldn't load on its own, including blob: URLs created by chrome
code. Full-text indexing loads HTML attachments through such a URL, so
indexing crashed Zotero with "Illegal load attempt of blob: URL from
web". Loads started by the parent are exempt, so start the load there
and use the child actor only to disable content retargeting.
https://forums.zotero.org/discussion/133661/
We only ever checked for corruption errors from transactions in
queryAsync(), so a corruption error raised by the COMMIT that mozStorage
runs itself bypassed the check. A 10.0 schema upgrade -- which heavily
exercises the database -- that failed due to corruption showed "Database
upgrade error" and a single Sqlite.sys.mjs frame instead of the prompt
offering to restore from a backup.
Two users reported this, but it's not clear what triggered it --
corruption usually occurs during a statement, which we did catch. There
may have been some statement transaction whose corruption error was
caught and ignored rather than being left to abort the transaction,
causing SQLite to then block the commit. That's what the test does, and
it fails without the fix.
https://forums.zotero.org/discussion/133611/
If the keystore's secret is replaced -- say, after a keyring reset --
the stored value won't be able to be decrypted again, so tell them to
try logging in again instead of trying to fix their keyring.
https://forums.zotero.org/discussion/133603/
unload() closes and re-adds the tab, and the close recorded an undo-close
history entry, so Undo Close Tab could open a duplicate of a tab that was
still open.
_addEntry() wrote the row with REPLACE INTO, listing only itemID and
data, so flag reverted to its default of 0. It's called for every match
on every update, not just new ones, so a retraction the user had hidden
came back as soon as the list changed and Zotero restarted.
A window is marked as closed before its unload handlers run, so the tab
events fired from ZoteroPane.destroy() warned about every observer that
the closing window hadn't torn down yet. The leak check now also requires
the window to be detached from its docShell, and it covers functions. It
caught an observer that outlived its window -- the sync-reminder API key
observer, which is now unregistered on teardown.
Selecting a collection doesn't wait for the tag selector to re-render,
so waitForTagSelector() could be resolved by the initial render instead
of the awaited change, and the assertions could run against a stale tag
list.
The cached retractions list is saved with the current data version from
the code, which we can bump to make clients discard their cached data
and refetch it. We've never actually done that, but if we did, it
would've hit a bug: init() would've started the refetch but still sent
the cached ETag, so if the list itself hadn't changed, the server
would've returned 304 and nothing would have been updated. The stale
version then would have survived, so the refetch would have been retried
on every startup.
The shims from 5697ee0af7 keep old plugins working for users, but
beta/dev/source builds should still throw to get the attention of
developers and (hopefully breakage-tolerant) beta users.
Plugins written for earlier versions call methods like
CollectionTree#getSelectedSearch(), which now throw. We did this
intentionally to make old code more obviously broken, even with single
selection, so developers would fix their code, but we're not yet
blocking plugins that illegitimately declared compatibility with a
future version, so some haven't been updated and are breaking Zotero.
With a single row selected, the pre-10 functions can technically
continue to work, so for now, restore them and just warn and name the
plugin instead of throwing. They still need to throw if multiple rows
are selected, since plugins that haven't been updated can't safely act
on a selection that might span collections or libraries.
https://forums.zotero.org/discussion/133565/
The row values were computed with `row.isItem && …`, so collection and
search rows in the trash got the boolean `false`, which the table then
rendered as the text "false".
https://forums.zotero.org/discussion/133560/
A stored-file path of 'storage:/' -- left behind by the 128 schema step,
which skipped paths with no basename -- triggered an
NS_ERROR_FILE_UNRECOGNIZED_PATH that aborted the whole
checkForUpdatedFiles() loop, so no files synced in the library. Skip an
attachment that throws instead of failing the whole library.
Also apply the setter's directory-path rule in getFilePath[Async]() to
avoid errors elsewhere.
https://forums.zotero.org/discussion/133523/synchronize-issue
Collections are now tracked as they're selected, added to, or dropped
on, and the five most recent usable targets are listed above the
full collection hierarchy by full path.
transformToDocument() needs a load group, which it takes from the source
document or from the window that created the XSLTProcessor. Neither of
those has existed since 0f2690eb75 in Zotero 8 stopped taking
XSLTProcessor from the hidden window, so the CSL 0.8 → 1.0 upgrade threw
NS_ERROR_FAILURE. In Word, this showed as "Zotero encountered an error
while updating your document."
https://forums.zotero.org/discussion/133543/
citeproc-rs is no longer maintained, and people who had enabled the
hidden pref were hitting errors.
This also drops the free() call on CSL engines, which only existed to
free the citeproc-rs wasm driver.
https://forums.zotero.org/discussion/133515/
Every failure shows "User canceled OS unlock entry" no matter what went
wrong. Test the store for the actual state, and give callers a message
describing what couldn't be accessed.
Both the keystore fallback prompt and the migration alert are reachable
from the credential read paths that a sync uses, so don't show if it's
an automatic sync.
Some Linux systems have no Secret Service running, and if users can't
change that (e.g., a managed system), storing an API key fails and login
never completes. Offer to store credentials unencrypted instead, and try
to encrypt them on a later read if the keystore becomes usable.
https://forums.zotero.org/discussion/133418/
The conditions menu had its own find-as-you-type, which
Utilities.Internal.addMenuFindAsYouType() now provides for any menu.
Matching walks the menu rather than a cached list of every condition, so
typing can no longer select a condition the menu doesn't offer --
Collection and Saved Search are removed when the search spans multiple
libraries.
An attachment or annotation condition is shown with a short label inside
its submenu, so pass the full name to match on, which is what the
menulist shows once the condition is selected.
The Collection value menu was a flat list of every collection in the
library, with subcollections set apart by an indent (or hyphens before
606d8f19ba). Build it with Utilities.Internal.createMenuForTarget()
instead, using the new 'filter' feature to limit to collections. The new
FAYT helper preserves matching on subcollections.
If a subcollection is selected, open the path down to the subcollection
when opening the menu. (As of fx153, macOS renders popups as native
menus, which can't be opened to a submenu, so revert to non-native menus
there.) A collection in a submenu can't be a menulist's selected item,
so the condition holds the value and sets the menulist's label and icon
itself, with the collection's path as a tooltip.
Items are added to the collection created for an import as they're
saved, before the imported collection hierarchy exists, so an item only
in a subcollection ended up in the top collection as well as its own.
Remove those once the hierarchy has been created.
When everything imported belongs to a single top-level collection, that
collection is used for the import rather than being nested inside a
collection named after the file.
https://forums.zotero.org/discussion/133174/
Exporting a collection included only its subcollections, so an item
directly in the collection came through with no collection at all, and
an item in both the collection and a subcollection came through in only
the subcollection.
Exporting a selection of multiple collections exported a flat list of
items with no collections at all.
We now include the selected collection(s), with one exception: if a
saved search is also selected, we export a flat item list, since a
search can't be exported as a collection.
Mozilla's OSKeyStore.encrypt() encodes the string as UTF-8 before
encrypting, but its decrypt() returns the decrypted bytes as a binary
string without decoding them, so a WebDAV password containing non-ASCII
characters came back mojibake and authentication failed.
https://forums.zotero.org/discussion/133465/problem-login-into-webdav-server-with-10-0-1
Operators can be typed as the Advanced Search shows them, which every
locale already translates, and new keyword messages cover the join
words, "no"/"has", the units of a relative date, and the range forms.
Each range form is given as an example with its two ends filled in, so
that a locale can say it its own way and each form keeps its own words
(e.g., no "between 1970 to 2000").
"year is between 1970 and 2000", "year:1970-2000", "1970..2000", and
"1970 to 2000" all match values within the range, inclusive of both
ends. Ends can be a year, a month ("added between 2024-02 and
2024-06"), a day ("date:2020-03-01..2020-03-15"), or a count ("number
of tags between 2 and 5").
* Fix js-ctypes-based symlinking on Linux by using `libc.so.6` instead of `libc.so` in `OS.File.unixSymlink()` and `Zotero.File.createSymlink()`
* Use that instead of `/bin/ln`, which doesn't exist on NixOS
* Replace `/bin/ln` with `Zotero.File.createSymlink()` in symlinked-database test
---------
Co-authored-by: Dan Stillman <dstillman@zotero.org>
FSEvents is backed by a per-volume journal that only local volumes
have. On a network mount the stream is created and started
successfully but never delivers events, so the watcher would report
that nothing had changed for as long as it was used. Check the volume
with statfs() and fall back to scanning.
Since f21e1b2d32, a full local file scan no longer runs periodically and
on every manual sync, so locally missed attachments stayed marked for
upload and were skipped as unavailable instead of being downloaded.
"Reset File Sync History" marked every attachment for upload, including
files that had never been downloaded, so the forced download check added
in 404fc41b88 found nothing to download.
Missing files are now marked for download when the upload queue is
filled, and in at-sync-time mode they're downloaded in the same sync.
The reset marks them for download directly, and downloads are no longer
skipped just because there were no remote storage changes.
https://forums.zotero.org/discussion/133414/
Any error while saving a note prompted the user to restart Zotero, even
a transaction timeout caused by a long-running operation elsewhere.
Nothing has been written when the wait times out, so retry, unless newer
note content has been handed to the editor in the meantime.
https://forums.zotero.org/discussion/133298/
Editing a note flagged it stale, and the background drain indexed it and
then ran an FTS5 'optimize' -- a single statement that rewrites the
content index and can hold the shared database for over a minute --
because the queue was empty again. Note saves waiting on the connection
hit the transaction timeout and told the user to restart Zotero.
Merge the index in bounded steps instead, so no statement runs long
enough to keep other queries waiting, and only after enough items have
been indexed to be worth it.
https://forums.zotero.org/discussion/133298/
Any Field expands to a generic 'field' condition, which the cross-level
code treated as matching only on top-level items. At an attachment
result level it therefore matched attachments whose parent item had the
value, instead of attachments with the value themselves. Since the
condition stands in for every field, it's now treated as matching at any
level those fields live at.
Zotero.SearchQuery turns a query like `by:smith after:2020 tag:"to read"
crispr` or `creator is smith and (tag is foo or bar)` into a
Zotero.Search, matching whatever text is left over using the current
search mode. Anything that doesn't look like a clause is free text, so a
DOI or a title with a colon in it is matched literally.
The search box syntax-highlights the parts of recognized conditions and
offers autocomplete for condition names, for the values of conditions
that have a fixed set of them (like item type), and for tags and
creators from the selected libraries.
Values without a parsable year are stored with a 0000-00-00 sort key,
and the guard meant to exclude them from comparisons checked the full
multipart value, which is always greater than '0000-00-00'.
Update the global schema to 45, resolve a field's date type through its
base-field mapping in ItemFields.isDate() (to cover priorityDate), and
convert stored values of date-type fields to multipart dates on schema
upgrade.
Only Date and Accessed were routed to the datefield condition, so other
date fields (e.g., Filing Date) offered only text operators and
couldn't be compared as dates. Text operators are still accepted on
date fields so existing saved searches keep loading and running.
zotero.exe is the Mozilla launcher process, which spawns the browser
process and exits, so the test runner checked for the success file and
cleaned up the profile while the app was still starting, and passing
runs were reported as failures that ended at arbitrary points.
CIFS mounts can't create real symlinks without special mount options,
so skip the symlinked-database and broken-symlink tests when a created
symlink doesn't exist or isn't visible as one.
On macOS, SQLite chooses locking methods based on the filesystem
containing the database, and network filesystems (e.g., SMB, NFS),
read-only volumes, and filesystems without byte-range locking get
methods without shared-memory support, which WAL requires. Opening a
database with an adjacent WAL file on those crashes -- Mozilla's VFS
wrapper hides the missing shared-memory methods from SQLite's WAL
support check -- so the first Zotero 10 run converted the database to
WAL and every launch after that segfaulted during connection
initialization.
On macOS, mirror SQLite's method selection and use a rollback journal
when shared memory isn't available, converting an existing WAL database
before opening it: an empty WAL by reverting the header format versions
in place, and a non-empty WAL by replaying it into a temporary copy on
local disk that replaces the database file only after passing an
integrity check. A WAL file next to an already-converted database
(e.g., from an interrupted conversion) goes through the same
conversion. Also use openNotExclusive during integrity checks and
corruption recovery, which otherwise fail on SMB shares with an I/O
error from the exclusive open lock.
This corrects 22055d92b7, which passed openNotExclusive on all
platforms for an open failure described as affecting macOS and Linux,
and expected locking_mode=EXCLUSIVE to keep the WAL index in heap
memory with no -shm file. Neither claim held up: mozStorage opens the
WAL while initializing the connection, before any pragma can run, so
the index uses shared memory unless the exclusive VFS is in use, and
the exclusive open works on Linux CIFS mounts -- where non-exclusive
access instead made SQLite's lock-upgrade sequence trip over the SMB
byte-range lock mapping, leaving startup hung and the database never
created. So pass openNotExclusive only on macOS. On Linux this restores
unix-excl, which performs all locking under a single held lock and
keeps the WAL index in heap memory; Windows has no distinct exclusive
VFS and is unaffected.
https://forums.zotero.org/discussion/133258/
The traversal waits on focus/blur events, which Gecko fires only while
the window is active, so a window that can't be activated -- another
app focused during a local run, or intermittent deactivation in CI --
timed out after 10 seconds instead of reporting that it couldn't run.
