* Add option to exclude fields from unqualified searches
* Use temp tables instead
This is slightly faster according to my (very rough) tests.
* Make query a bit more readable
* exclude_from_search -> excludeFromSearch
* Remove superfluous notetypes table from query
* Rework to use field search logic
Thanks to Rumo for the suggestion: https://github.com/ankitects/anki/pull/2394#issuecomment-1446702402
* Exclude fields from field searches too
* Fix error on notetypes with no included fields
* Add back the exclude_fields function
This approach seems to perform better on average than the previously
benchmarked ones.
* Use pure-SQL approach to excluding fields
* Change single field search to use new approach
* Fix flawed any_excluded/sortf_excluded logic
* Support field exclusion in the nc operator
Also fix search text being wrapped in % in the any_excluded=true case.
* Support field exclusion in the re and w operators
* Label field exclusion as being slower
* Unqualified search should be wrapped in % in all cases
I was under the impression that it shouldn't be wrapped with the new
field exclusion logic.
* Remove unnecessary .collect()
* Refactor some complex return types into structs
* Do not exclude fields in field searches
* Add a test and docstring for CollectRanges
* Avoid destructuring in closures
* Remove the exclude_fields function
Minor wording tweaks by dae:
* num_fields -> total_fields_in_note
* fields -> field_ranges_to_search
* fields -> fields_to_search
* SingleField -> FieldQualified
* mid -> ntid
* Ensure state mutator runs after card is rendered
* Ensure ease buttons only show when states are ready
* Pass context into states mutator
* Revert queuing of state mutator hook
Now that context data is exposed users shouldn't rely on the question
having been rendered anymore.
* Use callbacks instead of signals and timeout
... to track whether the states mutator ran or failed.
* Make mutator async
* Remove State enum
* Reduce requests and compute seed on backend
* List actions and locals in debug console
* Ignore whitespace when wrapping line with pp
* Scroll down after printing in debug console
Was previously preserving relative vertical position.
* Add feature to open and save debug scripts
* Refactor debug console into own module
* Add buffers to switch scripts
* Add action to delete script
* Revert "Fix open editors getting carried over to different notetypes (#2393)"
This reverts commit bf5bcd3f52.
* Improved fix for open editors getting carried over to different notetypes
* Run ninja format
Flask's .send_file() method sends a content-disposition header based
on the filename. If the filename includes non-Latin text, it adds an
rfc5987 unicode filename, but does not percent-escape the commas.
This causes Chromium to fail to load the image with the following
error:
net::ERR_RESPONSE_HEADERS_MULTIPLE_CONTENT_DISPOSITION
curl -v http://127.0.0.1:40000/Captura%20de%20ecr%C3%A3%202023-02-26,%20%C3%A0s%2018.33.03.png
* Trying 127.0.0.1:40000...
* Connected to 127.0.0.1 (127.0.0.1) port 40000 (#0)
> GET /Captura%20de%20ecr%C3%A3%202023-02-26,%20%C3%A0s%2018.33.03.png HTTP/1.1
> Host: 127.0.0.1:40000
> User-Agent: curl/7.86.0
> Accept: */*
>
* Mark bundle as not supporting multiuse
< HTTP/1.1 200 OK
< Access-Control-Allow-Origin: *
< Cache-Control: public, max-age=3600
< Content-Disposition: inline; filename="Captura de ecra 2023-02-26, as 18.33.03.png"; filename*=UTF-8''Captura%20de%20ecr%C3%A3%202023-02-26,%20%C3%A0s%2018.33.03.png
Fixed by supplying a dummy filename.
When a user clicks on the table row in the deck list (neither click on the deck
link nor on the options menu) while the shift key is pressed,
then set the deck in that table row as the current deck.
This gives the user some quick feedback and enables the user to easily create
subdecks (as it automatically fills the add deck dialog with the deck name of
the selected deck) and to create notes belonging to the selected deck.
* Remove Pane components and use Collapsible for TagEditor
* Update translations
* Give TagEditor border and focus outline
* Use ScrollArea from #2248 for fields
* Refactor ScrollArea
* Fix error caused by calling bridgeCommand when it's not available
* Make sure tag editor fills whole width of container
which is important for the CSV import page.
* Update NoteEditor.svelte
* Add back removed ftl strings
* Fix tests (dae)
* Implement TTS using windows crate
* Use API calls instead of SSML
* Properly stop player in case of TTS error
* Add context to WindowsErrors
* Validate available voices
* Remove TTS text from synthesize error
* Limit maximum buffer size
* Make validation optional and list it in tts filter
* We no longer need the winrt module (dae)
* Use a separate request object so the meaning of the bool is clear (dae)
* Slightly shorten runtime error message (dae)
The default message appears to clip slightly.
* Alternate buffer implementation (dae)
* Use array instead of vec
* Drop the max buffer size to 128k (dae)
* Replaced ankimedia object directly call by addon specific hook
# Conflicts:
# qt/aqt/browser/previewer.py
# qt/aqt/clayout.py
# qt/aqt/reviewer.py
* Replaced ankimedia.js by addon specific hook
# Conflicts:
# qt/aqt/browser/previewer.py
# qt/aqt/clayout.py
# qt/aqt/main.py
* Create specific location name for each hook to reuse control
* Created the card_review_webview_did_init hook
* Extended the hook card_will_show to replace will_show_web
The new hook card_will_show_state takes three new arguments
* Created the hook audio_did_pause_or_unpause to replace will_show_web
The new hook is called when audio toggle pause is called
* Created the hook audio_will_replay to replace will_show_web
The new hook is called when the audio is replayed by the user.
* Created the hook previewer_will_redraw_after_show_both_sides_toggled
to replace will_show_web.
The new hook fully replaces the last uses of will_show_web.
* Replaced card_will_show_state hook with reviewer_did_init and
equivalents. Instead of receiving the required state, it access it
by caching the object values with hooks as reviewer_did_init.
* Create common web view registry and unify title setting
* Consistently use space-separated naming for webview titles
None of the modified titles seem to be in use by add-ons, so we are not bound to the current naming.
The old naming was likely following camelCase as the name was also acting as a key for saveGeom, which is no longer the case.
* Update webview_did_inject_style_into_page example
* Add docstring to addon-targeted method
* Change AnkiWebView.origin to property
* Fix dupe enum value
* Tweak method name
* Add semicolon
* Rename `AnkiWebViewOrigin` to `AnkiWebViewKind`
* Allow burying cards in browser
This code is based on existing "toggle suspend" command in browser.
- Adds "toggle bury" command to browser cards menu.
- Adds "browsing-toggle-bury" to core translation. Only english-language.
- Adds "buried" coloring to rows for buried cards in browser table.
Not yet done:
- Keyboard shortcut for "toggle bury" action.
- Non-english translations.
* Add contributor as requested in CONTRIBUTORS.
* Fix formatting in browser_table.rs.
* Add keyboard shortcut to "toggle bury" command.
This adds keyboard shortcut "ctrl-shift-j" to "toggle bury" command in
browser cards menu.
* Simplify logic for color of buried-card rows.
On Linux distributions that do not yet support org.freedesktop.appearance.color-scheme (most distros released prior to 2022), querying DBus for the missing property produces verbose journal entries over Anki's entire runtime.
* Create widget gallery dialog
* Add WidgetGallery to debug dialog
* Use enum for its intended purpose
* Rename "reduced-motion" to "reduce-motion"
* Add another border-radius value
and make former large radius a bit smaller.
* Revamp preferences, add minimalist mode
Also:
- create additional and missing widget styles and tweak existing ones
- use single profile entry to set widget styles and reduce choices to Anki and Native
* Indent QTabBar style definitions
* Add missing styles for QPushButton states
* Fix QTableView background
* Remove unused layout from Preferences
* Fix QTabView focused tab style
* Highlight QCheckBox and QRadioButton when focused
* Fix toolbar styles
* Reorder preferences
* Add setting to hide bottom toolbar
* Move toolbar settings above minimalist modes
* Remove unused lines
* Implement proper full-screen mode
* Sort imports
* Tweak deck overview appearance in minimalist mode
* Undo TitledContainer changes
since nobody asked for that
* Remove dynamic toolbar background from minimalist mode
* Tweak buttons in minimalist mode
* Fix some issues
* Reduce theme check interval to 5s on Linux
* Increase hide timer interval to 2s
* Collapse toolbars with slight delay when moving to review state
This should ensure the bottom toolbar collapses too.
* Allow users to make hiding exclusive to full screen
* Rename full screen option
* Fix hide mode dropdown ignoring checkbox state on startup
* Fix typing issue
* Refine background image handling
Giving the toolbar body the main webview height ensures background-size: cover behaves exactly the same.
To prevent an override of other background properties, users are advised to only set background-images via the background-image property, not the background shorthand.
* Fix top toolbar getting huge when switching modes
The issue was caused by the min-height hack to align the background images. A call to web.adjustHeightToFit would set the toolbar to the same height as the main webview, as the function makes use of document.offsetHeight.
* Prevent scrollbar from appearing on bottom toolbar resize
* Cleanup
* Put review tab before editing; fix some tab orders
* Rename 'network' to 'syncing'
* Fix bottom toolbar disappearing on UI > 100
* Improve Preferences layout by adding vertical spacers to the bottom
also make the hiding of video_driver and its label more obvious in preferences.py.
* Fix bottom toolbar animating on startup
Also fix bottom toolbar not appearing when unchecking hide mode in reviewer.
* Hide/Show menubar in fullscreen mode along with toolbar
* Attempt to fix broken native theme on macOS
* Format
* Improve native theme on other systems by not forcing palette
with the caveat that theme switching can get weird.
* Fix theme switching in native style
* Remove redundant condition
* Add back check for Qt5 to prevent theme issues
* Add check for macOS before setting fusion theme
* Do not force scrollbar styles on macOS
* Remove all of that crazy theme logic
* Use canvas instead of button-bg for ColorRole.Button
* Make sure Anki style is always based on Fusion
otherwise we can't guarantee the same look on all systems.
* Explicitly apply default style when Anki style is not selected
This should fix the style not switching back after it was selected.
* Remove reduncant default_palette
* Revert 8af4c1cc2
On Mac with native theme, both Qt5 and Qt6 look correct already. On
the Anki theme, without this change, we get the fusion-style scrollbars
instead of the rounded ones.
* Rename AnkiStyles enum to WidgetStyle
* Fix theme switching shades on same theme
* Format
* Remove unused placeholderText
that caused an error when opening the widget gallery on Qt5.
* Check for full screen windowState using bitwise operator
to prevent error in Qt5.
Credit: https://stackoverflow.com/a/65425151
* Hide style option on Windows
also exclude native option from dropdown just in case.
* Format
* Minor naming tweak
This PR replaces the existing Python-driven sync server with a new one in Rust.
The new server supports both collection and media syncing, and is compatible
with both the new protocol mentioned below, and older clients. A setting has
been added to the preferences screen to point Anki to a local server, and a
similar setting is likely to come to AnkiMobile soon.
Documentation is available here: <https://docs.ankiweb.net/sync-server.html>
In addition to the new server and refactoring, this PR also makes changes to the
sync protocol. The existing sync protocol places payloads and metadata inside a
multipart POST body, which causes a few headaches:
- Legacy clients build the request in a non-deterministic order, meaning the
entire request needs to be scanned to extract the metadata.
- Reqwest's multipart API directly writes the multipart body, without exposing
the resulting stream to us, making it harder to track the progress of the
transfer. We've been relying on a patched version of reqwest for timeouts,
which is a pain to keep up to date.
To address these issues, the metadata is now sent in a HTTP header, with the
data payload sent directly in the body. Instead of the slower gzip, we now
use zstd. The old timeout handling code has been replaced with a new implementation
that wraps the request and response body streams to track progress, allowing us
to drop the git dependencies for reqwest, hyper-timeout and tokio-io-timeout.
The main other change to the protocol is that one-way syncs no longer need to
downgrade the collection to schema 11 prior to sending.
* Layout toolbar using CSS grid, introducing left and right trays
The trays provide a space for add-ons to introduce their own widgets to the toolbar without interfering with each other.
* Align tray items to the top
* Move absolutely positioned add-on items to right toolbar tray
Workaround that fixes breakages in add-ons like AMBOSS, Study Timer, and potentially others that currently still inject absolutely positioned elements into the toolbar using `top_toolbar_did_init_links`.
* Account for add-ons that add manual padding (e.g. Study Timer)
* Add docstrings and slightly refactor
* Tweak item alignment
* Introduce hooks for extending left and right toolbar trays
* Assign CSS classes to all tray items
* Add disclaimer on transitional nature of new hooks
* Give webviews a slide-in animation
if reduced motion isn't set.
* Auto-hide toolbar in review mode
moving the mouse above the main webview expands the toolbar. When the mouse leaves the toolbar, it will collapse after a delay of 2s.
* Save some space on bottom toolbars
* Use props for all hard-coded transition durations
and decrease most commonly used duration (200ms) to 150ms.
* Move auto-hide logic into ToolbarWebView
and handle auto-hide specific events in the respective webview subclasses.
* Fix typing issues
* Fix flickering issue
* Add auto_hide_toolbar opt-in to preferences
* Rename hide_toolbar to collapse_toolbar
to better describe the dock-like behaviour.
* Rename setting to minimize_distractions
* Reduce calls to pm in eventFilter
* Run formatter
* Revert setting title to something more specific
* Increase default animation time to 180ms
* Inset toolbar in review mode
when auto-hide is not enabled.
* Use card background on toolbar and add glass effect
* Use flatten/elevate over inset/outset
* Use flatten/elevate over inset/outset
* Update toolbar.py
* Fix toolbar background delay
* Tweak styles
* Use "collapse" instead of "auto-hide"
* Fix background misalignment in collapse mode
* Do not collapse toolbar when pointer is outside MainWebView
* Reduce hide_timer interval to 1000ms
* Use CSS to hide toolbar instead of setting webview height
* Add guard to prevent backdrop-filter: blur on Qt 5.14
* Apply transition to body instead of toolbar
to not complicate things for #2301.
* Fix Qt 5.14 and apply guard globally
* Fix background image scaling difference
* Tweak preference wording (dae)
Qt seems to be treating min-width more like width, and truncates text.
The issue is not limited to macOS with force_custom_styles: with a
sufficiently large amount of text on a button, it truncates on other
platforms too.
https://forums.ankiweb.net/t/2-1-55-issue-with-auto-browser-and-upload-download-dialog/25636
There are a few other uses of min-width in the styling - it may be worth
checking whether they suffer from the same issue as well.
The Rust community appear to have converged on tracing - it's used by
the Rust compiler, and receives close to 10x the number of downloads
that slog does. Its API is more ergonomic, and it does a much nicer
job with async rust.
To make this change, we no longer pass around explicit loggers, and rely
on a globally-registered one. The log file location has been changed
from one in each profile folder to a single one in the base folder. This
will remain empty for most users, since only errors are logged by default,
but may be useful for debugging future changes.
* Use a webview to show add-on's config help
This allows add-ons to embed images for example.
* Improve initial size of splitter widgets
* Decrease font size and margin of webview
This reverts commit 5dc79e22cd.
I appear to have been confused in my earlier testing, as reverting
this change seems to make no difference the top bar on first startup,
and fixes a bunch of other regressions that the original change introduced.
https://github.com/ankitects/anki/issues/2269
* Swap flag and mark indicator position in RTL mode
* Make buttons of bottom toolbar align to edge of screen in RTL mode
* Use start instead of left and end instead of right
Ensures the background is the correct color by the time the webview
is shown. We keep the #night check for now, as it's useful when testing
in an external browser.
The starting size of a webview seems to be 640x480, but if it is hidden
without retainSizeWhenHidden being set, the dialog it contains can end
up with a height of 0, which prevents the dialog from being shown.
By being explicit about our desired starting size, we can use a more
useful default, and avoid the issue of missing dialogs.
* Facilitate updating of hooks
- Add instructions in contributing.md
- Change addon_config_editor_will_update_json hook to work with the new
hookslib code
* Fix typo in docs
* Always run replaced hook
* Use lowercase list for typing
* Forbid defining both a replaced and a legacy hook
* Persist collapsed- and field states with SessionOptions object
* Format types.ts
* Replace format function with f-string
* Give setters more descriptive parameter names
* Do not use default prefix for descriptions and fonts
since they are not meant to be changed via Svelte.
When opening the graphs screen in dark mode, we want to load the
page first and then reveal the webview, to prevent a flash of white
that can appear as the page loads. Previously we did this for any
call to load_ts_page(), but this results in flicker when refreshing
an existing webview, such as the move from deck list to congrats screen.
In those cases, at least on the machines I have to test with here, the
refresh is smoother without the hide and show step.
The new window case is still not ideal - while the hide+show prevents a
flash of white, there is a flash of black instead, presumably as the
webview draws the initially-blank framebuffer with the contents of the
webview.
* Fix QMenu item not having different color on hover
due to the color changes in #2220.
* Remove strong border on pressed Qt widgets
* Make button gradient more subtle
by changing gradient-start on hover instead of gradient-end.
* Apply QPushButton style to QSpinBox buttons
* Improve margin of QComboBox arrow
* Make button-bg same color as button-gradient-end
This makes the hover gradient more subtle.
This reverts commit fa4fc3e15a.
Issue turned out to be a packaging problem, and this should not be
required as the socket should be held open even if removed.
This reverts commit ee70006ec4.
There have been a number of people complaining that the current
behaviour is not intuitive, and they have a point - it's not the
way browsers behave when you navigate between pages.
I was able to reproduce the crashes fairly reliably by opening the
prefs screen on startup and shutting down the app after 600ms; after
this change the crashes no longer seem to occur.
* Give deck browser table an elevated container look
* Tweak colors of elevated containers (e.g. in deck options)
* Prepare editor fields for custom backgrounds
* Tweak field margin and add explanatory comment
(for upgrading users, please see the notes at the bottom)
Bazel brought a lot of nice things to the table, such as rebuilds based on
content changes instead of modification times, caching of build products,
detection of incorrect build rules via a sandbox, and so on. Rewriting the build
in Bazel was also an opportunity to improve on the Makefile-based build we had
prior, which was pretty poor: most dependencies were external or not pinned, and
the build graph was poorly defined and mostly serialized. It was not uncommon
for fresh checkouts to fail due to floating dependencies, or for things to break
when trying to switch to an older commit.
For day-to-day development, I think Bazel served us reasonably well - we could
generally switch between branches while being confident that builds would be
correct and reasonably fast, and not require full rebuilds (except on Windows,
where the lack of a sandbox and the TS rules would cause build breakages when TS
files were renamed/removed).
Bazel achieves that reliability by defining rules for each programming language
that define how source files should be turned into outputs. For the rules to
work with Bazel's sandboxing approach, they often have to reimplement or
partially bypass the standard tools that each programming language provides. The
Rust rules call Rust's compiler directly for example, instead of using Cargo,
and the Python rules extract each PyPi package into a separate folder that gets
added to sys.path.
These separate language rules allow proper declaration of inputs and outputs,
and offer some advantages such as caching of build products and fine-grained
dependency installation. But they also bring some downsides:
- The rules don't always support use-cases/platforms that the standard language
tools do, meaning they need to be patched to be used. I've had to contribute a
number of patches to the Rust, Python and JS rules to unblock various issues.
- The dependencies we use with each language sometimes make assumptions that do
not hold in Bazel, meaning they either need to be pinned or patched, or the
language rules need to be adjusted to accommodate them.
I was hopeful that after the initial setup work, things would be relatively
smooth-sailing. Unfortunately, that has not proved to be the case. Things
frequently broke when dependencies or the language rules were updated, and I
began to get frustrated at the amount of Anki development time I was instead
spending on build system upkeep. It's now about 2 years since switching to
Bazel, and I think it's time to cut losses, and switch to something else that's
a better fit.
The new build system is based on a small build tool called Ninja, and some
custom Rust code in build/. This means that to build Anki, Bazel is no longer
required, but Ninja and Rust need to be installed on your system. Python and
Node toolchains are automatically downloaded like in Bazel.
This new build system should result in faster builds in some cases:
- Because we're using cargo to build now, Rust builds are able to take advantage
of pipelining and incremental debug builds, which we didn't have with Bazel.
It's also easier to override the default linker on Linux/macOS, which can
further improve speeds.
- External Rust crates are now built with opt=1, which improves performance
of debug builds.
- Esbuild is now used to transpile TypeScript, instead of invoking the TypeScript
compiler. This results in faster builds, by deferring typechecking to test/check
time, and by allowing more work to happen in parallel.
As an example of the differences, when testing with the mold linker on Linux,
adding a new message to tags.proto (which triggers a recompile of the bulk of
the Rust and TypeScript code) results in a compile that goes from about 22s on
Bazel to about 7s in the new system. With the standard linker, it's about 9s.
Some other changes of note:
- Our Rust workspace now uses cargo-hakari to ensure all packages agree on
available features, preventing unnecessary rebuilds.
- pylib/anki is now a PEP420 implicit namespace, avoiding the need to merge
source files and generated files into a single folder for running. By telling
VSCode about the extra search path, code completion now works with generated
files without needing to symlink them into the source folder.
- qt/aqt can't use PEP420 as it's difficult to get rid of aqt/__init__.py.
Instead, the generated files are now placed in a separate _aqt package that's
added to the path.
- ts/lib is now exposed as @tslib, so the source code and generated code can be
provided under the same namespace without a merging step.
- MyPy and PyLint are now invoked once for the entire codebase.
- dprint will be used to format TypeScript/json files in the future instead of
the slower prettier (currently turned off to avoid causing conflicts). It can
automatically defer to prettier when formatting Svelte files.
- svelte-check is now used for typechecking our Svelte code, which revealed a
few typing issues that went undetected with the old system.
- The Jest unit tests now work on Windows as well.
If you're upgrading from Bazel, updated usage instructions are in docs/development.md and docs/build.md. A summary of the changes:
- please remove node_modules and .bazel
- install rustup (https://rustup.rs/)
- install rsync if not already installed (on windows, use pacman - see docs/windows.md)
- install Ninja (unzip from https://github.com/ninja-build/ninja/releases/tag/v1.11.1 and
place on your path, or from your distro/homebrew if it's 1.10+)
- update .vscode/settings.json from .vscode.dist
