- note.flush() behaves like before, as otherwise actions or add-ons
that perform bulk flushing would end up creating an undo entry for
each note
- added col.update_note() to opt in to the new behaviour
- tidy up the names of some related routines
Reviews and operations on the backend that support undoing can now be
committed immediately, so they will not be lost in the event of a crash.
This required tweaks to a few places:
- don't set collection mtime on save() unless changes were made in
Python, as otherwise we end up accidentally clearing the backend undo
queue
- autosave() is now run on every reset()
- garbage collection now runs in a timer, instead of relying on
autosave() to be run periodically
- use dataclasses for the review/checkpoint undo cases, instead of the
nasty ad-hoc list structure
- expose backend review undo to Python, and hook it into GUI
- redo is not currently exposed on the GUI, and the backend can only
cope with reviews done by the new scheduler at the moment
- the initial undo prototype code was bumping mtime/usn on undo, but
that was not ideal, as it was breaking the queue handling which expected
the mtime to match. The original rationale for bumping mtime/usn was
to avoid problems with syncing, but various operations like removing
a revlog can't be synced anyway - so we just need to ensure we clear the
undo queue prior to syncing
- SearchTerm -> SearchNode
- Operator -> Joiner; share between messages
- build_search_string() supports specifying AND/OR as a convenience
- group_searches() makes it easier to negate
That way, the caller doesn't have to hold a reference to the browser and
explicitly call it again, if it wants to search for something specific.
Also, if the browser was closed and opened for a single-card-search, it
now won't perform a redundant current-deck-search first.
- IdList could be re-used for a cids: search in the future if required.
- Embedding the message means it's easy to access from Python as
an attribute of SearchTerm.
Manually hide the editor's completer popup before accepting a short cut which changes the focus widget. Thus, work around the bug in PyQt 5.15 where a QLineEdit with open QCompleter popup won't get notified about focus changes handled by the Main Window, leading to inconsistent focus.
Initially, I wanted to solve the bug reported on
https://github.com/Arthur-Milchior/anki-html-src-in-field/issues/1
After some research, I finally discovered that the trouble was that,
when we change the note type in add card, the method
`aqt.editor.Editor.loadNote` is called twice. In itself, it would not
be a problem, but given the way callback works, its call back is
called twice on the last version of the webview. Which means that
`gui_hooks.editor_did_load_note` is called twice, which breaks this
add-on.
The reason why loadNote is called twice is because `setNote` is called
twice in `aqt.modelchooser.onModelChange`. The first time through
`gui_hooks.current_note_type_did_change` which calls
`addcards.AddCards.onModelChange` which calls `loadNote`, the second
time through `self.mw.reset()` which calls
`gui_hooks.state_did_reset()` which calls `addcards.AddCards.onReset`
which calls `setAndFocusNote` which calls `setNote`.
I should note furthermore that currently,
`gui_hooks.current_note_type_did_change` is called only when the model
chooser change a model. And `addCards.onModelChange` is never called,
only added to the hook `gui_hooks.current_note_type_did_change`. So
removing the line of code removed in this commit will have no side
effect in Anki itself. It will only affect the fact that this method
is called twice.
I do not know of any add-on calling `onModelChange` or
`gui_hooks.current_note_type_did_change`, but it means little, so of
course, it may always mean an add-on will break because of this
change. No way of being sure.
The type hints allow mypy to check the gui_hook calls, revealing a
bunch of places that are broken as they expect no arguments like the
legacy hooks.
To make mypy happy about PyQt's signal.connect(func), a qconnect()
helper has been added.
