Some add-ons fully override AnkiWebPage.__init__ and thus depend on _setupBridge injecting the JS bridge script.
With this change we account for these cases, while giving add-ons the opportunity to look for solutions that do not require overriding AnkiWebPage.__init__ completely.
We were previously inspecting the referrer, but that is spoofable,
and doesn't guard against other processes on the machine.
To accomplish this, we use a request interceptor to automatically
add an auth token to webviews with the right context. Some related
changes were required:
- We avoid storing _page, which was leading to leaks & warning on exit
- At webview creation (or set_kind() invocation), we assign either
an authenticated or unauthenticated web profile.
- Some of our screens initialize the AnkiWebView when calling, e.g.,
aqt.forms.stats.Ui_Dialog(). They then immediately call .set_kind().
This reveals a race condition in our DOM handling code: the webview
initialization creates an empty page with the injected script, which
causes a domDone signal to be sent back. This signal arrives after
we've created another page with .set_kind(), causing our code to think
the DOM is ready when it's not. Then when we try to inject the dynamic
styling, we get an error, as the DOM is not ready yet. In the absence
of better solutions, I've added a hack to set_kind() to deal with this
for now.
* fix: correct typo and adjust indentation in docstring
Fixed a small typo in the webview_did_inject_style_into_page docstring and adjusted indentation for consistency.
* Update CONTRIBUTORS
* Show "and others" at the end of the contributor list in the about dialog
* Make about addendum translatable
* Fix CONTRIBUTORS
* Fix CONTRIBUTORS
* Update ftl/qt/about.ftl (dae)
* Remove unused import
* Nit-pick on code comment
* Enable View Page/Config buttons only when 1 add-on selected
* Enable Cmd+W shortcut (on Mac only) to close Add-ons dialog
* Feat/grade now
* pass ci
* fix from_queue
* Refactor card answering to support from_queue flag
- Add `from_queue` field to `CardAnswer` struct and proto message
- Modify `answer_card_inner` to handle queue updates based on `from_queue`
- Remove `grade_card` method and consolidate card answering logic
- Update related test cases to set `from_queue` flag
* fix current_changes() called when no op set
* Optimize queue updates for batch card processing
- Refactor `grade_now` to collect processed card IDs first
- Add new `update_queues_for_processed_cards` method for efficient batch queue updates
- Improve queue management by removing entries and updating counts in a single pass
- Remove individual queue update method in favor of batch processing
* pass ci
* keep the same style
* remove ineffective code
* remove unused imports
* Add myself to CONTRIBUTORS
* Set draggable="false" attribute on .replay-button
Because currently if a user drags slightly (even unintentionally) upon clicking a play button, play does not happen
* Prevent dragging hint links
Because if a user moves cursor a little after `mousedown`, action (expanding the hint) does not occur. Which might cause issues from accessibility standpoint
* feat: add title to rename dialog
* fix: localize hardcoded message
* feat: add title to create deck dialog
* refactor: formatting fixes
* add name to about screen
adding my name to contributors for the third time
* Raise exception from closures in run_in_background using run_on_main
* Add author to CONTRIBUTORS file
* Undo taskman changes
* Raise exceptions from _on_closures_pending using singleShot
* refactor: accept window title in some dialog methods
* feat: match sync initial title with state
Sync iniates in the `Checked` state, so the initial title is changed to match it instead of using the generic `Anki` title
* feat: use `Sync` as the tile of login screen
Maybe a new string should be created for that.
* feat: Use `Sync` as title of the sync conflict dialog
Maybe a new string should be used for that
* refactor: fix formatting
* fix: alias in CONTRIBUTORS
Even in 2025, the script isn't smart enough to handle different casing or use just the GitHub ID
The uninstall script runs `xdg-mime uninstall` which takes >5 seconds to process.
There is no indication to the user though that the script is actually running. Adding an `echo` info message solves that.
Additionally we could `rm -rfv` to make it more verbose (the install script is verbose too). But the main thing that needs time to process is the `xdg-mime uninstall` part of the script.
The reason why I didn't make `rm -rf` verbose, too, is that the output text is greater than the buffer that the terminal provides – meaning you cannot view it from the beginning. And since `rm` is very fast even on old systems with slow hardware I didn't really see a reason to make it verbose here.
* Add function to restore the default name of a flag
* Call function to restore default flag name if flag renamed to empty string
* Update _load_flags to use the default_flag_names dict
* Add name to contributors file
* Add trailing comma to pass tests
* Update to follow python style guide
* Update about.py
* Revert "Update _load_flags to use the default_flag_names dict"
This reverts commit caa8fea94b.
* Use require_refresh() instead of storing default flag names
* fix discard changes randomly being a noop on certain linux systems
* use QApplication.setActiveWindow instead
* revert current impl
* wait for the next event loop iter before calling activeWindow
* allow adding images via drag/drop when adding io
* support editing io notes as well
Co-authored-by: Abdo <abdo@abdnh.net>
---------
Co-authored-by: Abdo <abdo@abdnh.net>
This reverts commit 04228de666.
Anki 25.01 Beta 1 revealed a bunch of regressions with the latest
Qt, and zero reports of improvements from it, so we'll be better off
holding off on it for now, and perhaps reporting the deadkeys issue
to Qt once we've got a proper reproduction process.
* Prevent stale frames being drawn.
At key points where external changes enter the webview, stale images might get rendered. This ensures that a frame showing current state is always shown.
* Only stage single redraw
* Remove potentially superfluous calls to `self.update()`
* Remove potentially superfluous calls to `self.update()`.
I lost this one during some git troubles.
* Revert unrelated change
The function is supposed to take a boolean telling it whether or not the loading succeeded, which it doesn't as is. However, this is unrelated and works either way so I also reverted it again.
* chore: code cleanup
* cleanup: Remove redundant check for presence of callback
A callback will be used either way for this call, so it can be simplified. The check happens inside the handler.
* Add comment explaining why this change is necessary, referencing the relevant PR.
* Clarify comment to answer the why, not the what.
One can see what is being done, why is probably more important.
* expose get_image_occlusion_fields
* fix create copy for io
* revert current impl
* passthru original note id when creating copy
* add IOCloningMode
* fix create copy for io
