* bump nom to 8.0.0
* update cloze.rs
* update template.rs
* update imageocclusion.rs
* update search/parser.rs
* update card_rendering/parser.rs
* replace use of fold_many0 with many0
in nom 8, `many0` doesn't accumulate when used within `recognize`
* Migrate build system to uv
Closes#3787, and is a step towards #3081 and #4022
This change breaks our PyOxidizer bundling process. While we probably
could update it to work with the new venvs & lockfile, my intention
is to use this as a base to try out a uv-based packager/installer.
Some notes about the changes:
- Use uv for python download + venv installation
- Drop python/requirements* in favour of pyproject files / uv.lock
- Bumped to latest Python 3.9 version. The move to 3.13 should be
a fairly trivial change when we're ready.
- Dropped the old write_wheel.py in favour of uv/hatchling. This has
the unfortunate side-effect of dropping leading zeros in our wheels,
which we could try hack around in the future.
- Switch to Qt 6.7 for the dev repo, as it's the first PyQt version
with a Linux/ARM WebEngine wheel.
- Unified our macOS deployment target with minimum required for ARM.
- Dropped unused fluent python files
- Dropped unused python license generation
- Dropped helpers to run under Qt 5, as our wheels were already
requiring Qt 6 to install.
* Build action to create universal uv binary
* Drop some PyOxidizer-related files
* Use Windows ARM64 cargo/node binaries during build
We can't provide ARM64 wheels to users yet due to #4079, but we can
at least speed up the build.
The rustls -> native-tls change on Windows is because ring requires
clang to compile for ARM64, and I figured it's best to keep our Windows
deps consistent. We already built the wheels with native-tls.
* Make libankihelper a universal library
We were shipping a single arch library in a purelib, leading to
breakages when running on a different platform.
* Use Python wheel for mpv/lame on Windows/Mac
This is convenient, but suboptimal on a Mac at the moment. The first
run of mpv will take a number of seconds for security checks to run,
and our mpv code ends up timing out, repeating the process each time.
Our installer stub will need to invoke mpv once first to get it validated.
We could address this by distributing the audio with the installer/stub,
or perhaps by putting the binaries in a .pkg file that's notarized+stapled
and then included in the wheel.
* Add some helper scripts to build a fully-locked wheel
* Initial macOS launcher prototype
* Add a hidden env var to preload our libs and audio helpers on macOS
* qt/bundle -> qt/launcher
- remove more of the old bundling code
- handle app icon
* Fat binary, notarization & dmg
* Publish wheels on testpypi for testing
* Use our Python pin for the launcher too
* Python cleanups
* Extend launcher to other platforms + more
- Switch to Qt 6.8 for repo default, as 6.7 depends on an older
libwebp/tiff which is unavailable on newer installs
- Drop tools/mac-x86, as we no longer need to test against Qt 5
- Add flags to cross compile wheels on Mac and Linux
- Bump glibc target to 2_36, building on Debian Stable
- Increase mpv timeout on macOS to allow for initial gatekeeper checks
- Ship both arm64 and amd64 uv on Linux, with a bash stub to pick
the appropriate arch.
* Fix pylint on Linux
* Fix failure to run from /usr/local/bin
* Remove remaining pyoxidizer refs, and clean up duplicate release folder
* Rust dep updates
- Rust 1.87 for now (1.88 due out in around a week)
- Nom looks involved, so I left it for now
- prost-reflect depends on a new prost version that got yanked
* Python 3.13 + dep updates
Updated protoc binaries + add helper in order to try fix build breakage.
Ended up being due to an AI-generated update to pip-system-certs that
was not reviewed carefully enough:
https://gitlab.com/alelec/pip-system-certs/-/issues/36
The updated mypy/black needed some tweaks to our files.
* Windows compilation fixes
* Automatically run Anki after installing on Windows
* Touch pyproject.toml upon install, so we check for updates
* Update Python deps
- urllib3 for CVE
- pip-system-certs got fixed
- markdown/pytest also updated
* modify render_card to return whether card was empty
* plumbing
* add flag to proto message
* plumbing: pass flag along to PartiallyRenderedCard
* add tests
* Use a custom return type for clarity (dae)
* typeanswer: fix cleanup
Fix: Add prepare_expected back in for the 'nothing typed' & 'correctly typed' cases. This also makes expected_original redundant again.
Style: %s/provided/typed/g
Style: rename one ch → c
Testcase: whitespace_is_trimmed: added a check for the "correctly typed" path and renamed it to tags_removed (there's no whitespace?)
Testcase: empty_input_shows_as_code: changed to also check that tags get trimmed
* [type:nc] – ignores combining characters
Adds a comparison variant to [type] which ignores when combining characters of the expected field are missing from the provided input. It still shows these characters in the 'expected' line for reference.
It's useful for languages with e.g. diacritics that are required for reference (such as in dictionaries), but rarely actually learned or used in everyday writing. Among these languages: Arabic, Hebrew, Persian, Urdu.
The bool 'combining' controls it as new final parameter of both relevant compare_answer functions. On the Python side, it's set to true by default.
Use on the note templates: [type:nc:field] (only the front needs to include :nc)
This also removes the need to have both variants of words/sentences present as separate fields, to show them redundantly, etc.
* typeanswer: simplify by using nfkd throughout
Requires adjusting two testcases, but both render exactly the same in Anki itself.
On NFC vs. NKFD: https://stackoverflow.com/a/77432079
* typeanswer: 'simplify' by removing normalize_typed (requiring a bool parameter)
I'd prefer to keep this extra method.
* typeanswer: micro-optimize vectors
Should get rid of most relocations, at the expense of over-allocating.
On Vec's (String's) behavior: https://stackoverflow.com/a/72787776
* Mark `combining` as private
typeCorrect is not marked as private either, but we can at least do
the right thing for newly-added code.
* Revert "typeanswer: micro-optimize vectors"
This reverts commit 9fbacbfd19.
* Revert "typeanswer: 'simplify' by removing normalize_typed (requiring a bool parameter)"
This reverts commit df2dd3394e.
Instead of flattening the output (which was missing FrontSide), alter
the behaviour of render() instead. The non-partial output is now exposed
via Protobuf, so the non-Python clients can take advantage of it.
* Automatically elide empty inputs and outputs to backend methods
* Refactor service generation
Despite the fact that the majority of our Protobuf service methods require
an open collection, they were not accessible with just a Collection
object. To access the methods (e.g. because we haven't gotten around to
exposing the correct API in Collection yet), you had to wrap the collection
in a Backend object, and pay a mutex-acquisition cost for each call, even
if you have exclusive access to the object.
This commit migrates the majority of service methods to the Collection, so
they can now be used directly, and improves the ergonomics a bit at the
same time.
The approach taken:
- The service generation now happens in rslib instead of anki_proto, which
avoids the need for trait constraints and associated types.
- Service methods are assumed to be collection-based by default. Instead of
implementing the service on Backend, we now implement it on Collection, which
means our methods no longer need to use self.with_col(...).
- We automatically generate methods in Backend which use self.with_col() to
delegate to the Collection method.
- For methods that are only appropriate for the backend, we add a flag in
the .proto file. The codegen uses this flag to write the method into a
BackendFooService instead of FooService, which the backend implements.
- The flag can also allows us to define separate implementations for collection
and backend, so we can e.g. skip the collection mutex in the i18n service
while also providing the service on a collection.
Due to the orphan rule, this meant removing our usages of impl ProtoStruct,
or converting them to a trait when they were used commonly.
rslib now directly references anki_proto and anki_i18n, instead of
'pub use'-ing them, and we can put the generated files back in OUT_DIR.
* 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)
* Run cargo +nightly fmt
* Latest prost-build includes clippy workaround
* Tweak Rust protobuf imports
- Avoid use of stringify!(), as JetBrains editors get confused by it
- Stop merging all protobuf symbols into a single namespace
* Remove some unnecessary qualifications
Found via IntelliJ lint
* Migrate some asserts to assert_eq/ne
* Remove mention of node_modules exclusion
This no longer seems to be necessary after migrating away from Bazel,
and excluding it means TS/Svelte files can't be edited properly.
(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
* Add new `card_rendering` mod
Parses a text with av/tts tags and strips or extracts tags.
* Replace old `extract_av_tags` and `strip_av_tags`
... with new `card_rendering` mod
* ressource -> resource
* Add AV prettifier for use in browser table
* Accept String in av tag routines
... and avoid redundant writes if no changes need to be made.
* add benchmarking with criterion; make links test optional (dae)
cargo install cargo-criterion, then run ./bench.sh
* performance comparison: creating HashMap up front (dae)
the previous solution:
anki_tag_parse time: [1.8401 us 1.8437 us 1.8476 us]
this solution:
anki_tag_parse time: [2.2420 us 2.2447 us 2.2477 us]
change: [+21.477% +21.770% +22.066%] (p = 0.00 < 0.05)
Performance has regressed.
* Revert "performance comparison: creating HashMap up front" (dae)
This reverts commit f19126a2f1.
* add missing header
* Write error message if tts lang is missing
* `Tag` -> `Directive`
