(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
* Reduce font size of answer button indicators
* Increase padding of browser rows with ResizeToContents on vertical header
* Remove 0.8 scale factor for dropdown item font-size
* Remove font-size prop entirely from DropdownItem
* Revert "Remove font-size prop entirely from DropdownItem"
This reverts commit bb0a158f96183cca74e198867070c2f99af04dc4.
* Remove hard-coded Python font sizes
* Move font size and scrollbar into _root-vars.scss
* Revert editor size variable to 1.6
* Fix icon alignment
* Fix checkbox alignment for dropdown items
* Remove unused classes from Tag.svelte
* Revert "Increase padding of browser rows with ResizeToContents on vertical header"
This reverts commit 77bfc854ba140dd99aae98efcdd4af7052615fa6.
* Remove option to set font size of browser entries
* Add setting for browser row padding to preferences
* Revert "Add setting for browser row padding to preferences"
This reverts commit 75c59da65a1028e2caa3c48b247f99825c1b0b6c.
* Revert "Remove option to set font size of browser entries"
This reverts commit a543783d8ea079f39b7ae445152573c96be29841.
* Prevent multiple inclusion of variables in CSS files
* Use dict instead of tuple for variables
* Add comments to variables
* Improve appearance of main window
* Tweak main window styles
* Use json.dumps over pprint.format
* Make study button primary
* Improve header margin
* Make bottom toolbar slimmer
* Make congrats page more balanced
* Fix type issue
* Replace day/night with light/dark
* Exclude top-level-drag-row from hover effect
* Create dataclass for variables
* Run formatter
* Apply CSS variables from Python side
Why go full-circle with the Sass variables? This way we only need one interface for add-on authors to interact with. It also makes it easier for us to apply additional themes in the future.
* Fix typing
* Fix rgba values in Qt
* Darken button background
* Fix palette not being applied in light theme
For some odd reason this problem arose much later than #2016.
* Tweak default button look
* Reformat
* Apply CSS vars to ts pages
* Include elevation in button_mixins_lib
* Cast opacity to int
* Add some margin to studiedToday info
* Tweak light theme button gradient
* Tweak highlight-bg for light theme
* Add back default button color
as it made the browser sidebar tool icons dark in light theme.
* Reformat
* Tweak light theme buttons once more
Sorry for the back-and-forth. Sass only compiles when there are changes in user files, not when I only change the vars.
* Fix bottom toolbar button indicators
* Make buttons more clicky
* Fix button padding
* Handle macOS separately again
* Decrease elevation effect for main window buttons to 1
* Imitate box-shadow for Qt elements
* Adjust shadow vars
* Adjust primary border color
because the save button in the deck options had a lighter color than its background gradient.
* Boost box-shadow color of primary buttons
* Format
* Adjust Qt box-shadow imitation and shadow colors
* Use more subtle default shadow color
* Add some more padding to top toolbar
* Revert "Apply CSS vars to ts pages"
This reverts commit 5d8e7f6b7ffc8894b6517ecbb8cfba35407fc69a.
* Revert "Apply CSS variables from Python side"
This reverts commit 87db774412fd2bfd75e2630d2c5e782daef96b5f.
* Better match the standard macOS buttons
In the dark theme the standard color is a lighter grey, but at least
the size/shape is similar again.
This doesn't work for the editor buttons.
* Reduce the top margin of the congrats screen
* Fix illegible buttons when changing theme on macOS; match dark button style
* Remove --medium-border variable
* Implement color palette using Sass maps
I hand-picked the gray tones, the other colors are from the Tailwind CSS v3 palette.
Significant changes:
- light theme is brighter
- dark theme is darker
- borders are softer
I also deleted some platform- and night-mode-specific code.
* Use custom colors for note view switch
* Use same placeholder color for all inputs
* Skew color palette for more dark values
by removing gray[3], which wasn't used anywhere. Slight adjustments were made to the darker tones.
* Adjust frame- window- and border colors
* Give deck browser entries --frame-bg as background color
* Define styling for QComboBox and QLineEdit globally
* Experiment with CSS filter for inline-colors
Inside darker inputs, some colors like dark blue will be hard to read, so we could try to improve text-color contrast with global adjustments depending on the theme.
* Use different map structure for _vars.scss
after @hgiesel's idea: https://github.com/ankitects/anki/pull/2016#discussion_r947087871
* Move custom QLineEdit styles out of searchbar.py
* Merge branch 'main' into color-palette
* Revert QComboBox stylesheet override
* Align gray color palette more with macOS
* Adjust light theme
* Use --slightly-grey-text for options tab color
* Replace gray tones with more neutral values
* Improve categorization of global colors
by renaming almost all of them and sorting them into separate maps.
* Saturate highlight-bg in light theme
* Tweak gray tones
* Adjust box-shadow of EditingArea to make fields look inset
* Add Sass functions to access color palette and semantic variables
in response to https://github.com/ankitects/anki/pull/2016#issuecomment-1220571076
* Showcase use of access functions in several locations
@hgiesel in buttons.scss I access the color palette directly. Is this what you meant by "... keep it local to the component, and possibly make it global at a later time ..."?
* Fix focus box shadow transition and remove default shadow for a cleaner look
I couldn't quite get the inset look the way I wanted, because inset box-shadows do not respect the border radius, therefore causing aliasing.
* Tweak light theme border and shadow colors
* Add functions and colors to base_lib
* Add vars_lib as dependency to base_lib and button_mixins_lib
* Improve uses of default-themed variables
* Use old --frame-bg color and use darker tone for canvas-default
* Return CSS var by default and add palette-of function for raw value
* Showcase use of palette-of function
The #{...} syntax is required only because the use cases are CSS var definitions. In other cases a simple palette-of(keyword, theme) would suffice.
* Light theme: decrease brightness of canvas-default and adjust fg-default
* Use canvas-inset variable for switch knob
* Adjust light theme
* Add back box-shadow to EditingArea
* Light theme: darken background and flatten transition
also set hue and saturation of gray-8 to 0 (like all the other grays).
* Reduce flag colors to single default value
* Tweak card/note accent colors
* Experiment with inset look for fields again
Is this too dark in night mode? It's the same color used for all other text inputs.
* Dark theme: make border-default one shade darker
* Tweak inset shadow color
* Dark theme: make border-faint darker than canvas-default
meaning two shades darker than it currently was.
* Fix PlainTextInput not expanding
* Dark theme: use less saturated flag colors
* Adjust gray tones
* Fix nested variables not getting extracted correctly
* Rename canvas-outset to canvas-elevated
* Light theme: darken canvas-default
* Make canvas-elevated a bit darker
* Rename variables and use them in various components
* Refactor button mixins
* Remove fusion vars from Anki
* Adjust button gradients
* Refactor button mixins
* Fix deck browser table td background color
* Use color function in buttons.scss
* Rework QTabWidget stylesheet
* Fix crash on browser open
* Perfect QTableView header
* Fix bottom toolbar button gradient
* Fix focus outline of bottom toolbar buttons
* Fix custom webview scrollbar
* Fix uses of vars in various webviews
The command @use vars as * lead to repeated inclusion of the CSS vars.
* Enable primary button color with mixin
* Run prettier
* Fix Python code style issues
* Tweak colors
* Lighten scrollbar shades in light theme
* Fix code style issues caused by merge
* Fix harsh border color in editor
caused by leftover --medium-border variables, probably introduced with a merge commit.
* Compile Sass before extracting Python colors/props
This means the Python side doesn't need to worry about the map structure and Sass functions, just copy the output CSS values.
* Desaturate primary button colors by 10%
* Convert accidentally capitalized variable names to lowercase
* Simplify color definitions with qcolor function
* Remove default border-focus variable
* Remove redundant colon
* Apply custom scrollbar CSS only on Windows and Linux
* Make border-subtle color brighter than background in dark theme
* Make border-subtle color a shade brighter in light theme
* Use border-subtle for NoteEditor and EditorToolbar border
* Small patches
This was inadvertently dropped in 478b3a53f1,
but AnkiWeb requires it.
In the future we will need to give these index pages a bit more thought,
as it's not very helpful to be exporting a creation function if we're
automatically mounting as the module is imported. Dropping the automatic
mounting would give us more flexibility, but currently that leads to
a FOUC in night mode, as the webview is revealed prior to a subsequent
web.eval() call being able to fire.
* Make eslint sort our imports
* fix missing deps in eslint rule (dae)
Caught on Linux due to the stricter sandboxing
* Remove exports-last eslint rule (for now?)
* Adjust browserslist settings
- We use ResizeObserver which is not supported in browsers like KaiOS,
Baidu or Android UC
* Raise minimum iOS version 13.4
- It's the first version that supports ResizeObserver
* Apply new eslint rules to sort imports
* Add _bytes methods for all methods in the backend
Expose get_note in qt/aqt/mediasrv.py
* Satisfy formatter
* Rename _bytes function to _raw and have them bytes as input
* Fix backend generation
* Use lib/proto/deckOptions in deck-options
* Add exposed_backend to qt/aqt/mediasrv.py
* Move some more backend methods to exposed_backend_list
* Use protobufjs for congrats and i18n
* Use protobufjs for completeTag
* Use protobufjs services in change-notetype
* Reorder post handlers in alphabetical manner
* Satisfy tests
* Remove unused collection methods
* Rename access_backend to raw_backend_request
* Use _vendor.stringcase instead of creating a new function
* Remove SKIP_UNROLL_OUTPUT
* Directly call _run_command in non _raw methods
* Remove TranslateString, ChangeNotetype and CompleteTag from SKIP_UNROLL_INPUT
* Remove UpdateDeckConfigs from SKIP_UNROLL_INPUT
* Remove ChangeNotetype from SKIP_UNROLL_INPUT
* Remove SKIP_UNROLL_INPUT
* Fix typing issue with translate_string
- Adds typing support for Protobuf maps in genbackend.py
* Do not emit convenience method for protobuf TranslateString
* Move some AddCards specific code to NoteCreator.svelte
* Add new strings for Toggling the Visual / HTML editor
* Set LabelContainer vertical-align to text-top
- Makes them look more centered
* Remove appendInParentheses helper
* Make all ts/*.html files include only module.js and module.css
* Move any JS from .html to index files
* Remove .html files from ts modules
* Remove Python with Starlark implemenation
* Remove reference to non-existing file
* Remove deck-option.html as well
* fix change-notetype screen (dae)
* Refactor editor css, fix editor button highlight
- Avoid using webview.css
- Move more buttons css into button_mixins
* Fix DropdownItem appearance
* Fix the visuals of tags
* Make dropdown font slightly smaller
* Give SelectOption a background color
* Move some css from deck-options-base to CardStateCustomizer
* Avoid using core.scss for CardStats
* Avoid using sass/core in congrats package
* Inline core.scss into webview.scss
* Include fusion-vars for base.scss
* need to keep core.scss around for now (dae)
1. All Svelte files in a package are compiled in one step now, which
ensures that properties that use types from a different Svelte file in the
same package are typed correctly. The single-file svelte() has been removed,
and compile_svelte() may be renamed to svelte() in the future.
2. The .ts files in the same package are included as part of the Svelte
compilation, so that types imported imported from .ts files in the
same package work.
3. Dependencies passed into the rule are now loaded into the TypeScript
compiler, so that properties referencing types from different packages
work. We'll need to update our compile_svelte() lines to list the
dependencies. For example, before this change:
% cat bazel-bin/ts/congrats/CongratsPage.svelte.d.ts
import { SvelteComponentTyped } from "svelte";
declare const __propDef: {
props: {
info: any;
};
...
After adding //ts/lib to the deps of compile_svelte() in ts/congrats:
% cat bazel-bin/ts/congrats/CongratsPage.svelte.d.ts
import { SvelteComponentTyped } from "svelte";
import type { Scheduler } from "../lib/proto";
declare const __propDef: {
props: {
info: Scheduler.CongratsInfoResponse;
};
...
Merging note: the typing changes were fixed in a separate PR.
* Put rootDirs into subprojects
- typings do not work for any ts or svelte files
- if we set the 'rootDirs' in ts/tsconfig.json to '../bazel-bin/ts' and then inherit
them from e.g. editor, the root will be changed to '../../bazel-bin/ts',
however editor needs look in '../../bazel-bin/ts/editor' instead.
* Rename i18n and i18n_helpers to i18n-generated and i18n
- This way, we can restrict the awkwardness of importing files outside
the ts directory within lib
* Fix missing typing of i18n and backend_proto by adding back symlinks
* Split up i18n-generated into i18n-{translate,modules}
* Change i18n from singleton to functions
* Revert "Put rootDirs into subprojects"
This partially reverts commit e1d4292ce3979e7b7ee21bf3951b8a462d45c29c.
It seems like this might not be necessary after all.
However some other change made on this branch seems to have fixed
the .svelte.d.ts imports
* Introduce i18n-bundles to remove circular import
There was a circular import i18n.ts <-> i18n-translate.ts
* Create own directory for i18n
* Move lib/i18n/translate to lib/translate
* This restores tree shaking
* Update tsconfig libs and module
* es2018-2020 have wide support on all modern browsers including
* Switch bundles and langs inside i18n to variables again
* Add missing copyright header
* Rename translate.ts to ftl.ts
* Remove the symlinks again
I added them to fix to have completion for tr, however this would have
also have meant to abandon the tree shaking.
As we want to have tree shaking, it's also not necessary to have the
symlinks anymore
* Revert "Update tsconfig libs and module"
This reverts commit 0a96776a475e9901c1f9f3407c726d1d002fb9ef.
* move withCollapsedWhitespace back to i18n/utils
* Add back /ts as in rootDirs
The nice thing about the symlink approach is that it allowed tsc -b
to function without any changes to the tsconfig.json file, but it meant
there were extra links we had to maintain. So instead, we just add an
extra rootDirs entry, and add two commented-out lines that can be
uncommented when wanting to build with tsc directly.
ts_library() is deprecated and will presumably be dropped from a
future rules_nodejs, and it wasn't working with the jest tests
after updating, so we switch over to ts_project().
There are some downsides:
- It's a bit slower, as the worker mode doesn't appear to function
at the moment.
- Getting it working with a mix of source files and generated files
was quite tricky, especially as things behave differently on Windows,
and differently when editing with VS Code. Solved with a small patch
to the rules, and a wrapper script that copies everything into the
bin folder first. To keep VS Code working correctly as well, the built
files are symlinked into the source folder.
- TS libraries are not implicitly linked to node_modules, so they
can't be imported with an absolute name like "lib/proto" - we need
to use relative paths like "../lib/proto" instead. Adjusting "paths"
in tsconfig.json makes it work for TS compilation, but then it fails
at the esbuild stage. We could resolve it by wrapping the TS
libraries in a subsequent js_library() call, but that has the downside
of losing the transient dependencies, meaning they need to be listed
again. Alternatively we might be able to solve it in the future by
adjusting esbuild, but for now the paths have been made relative to
keep things simple.
Upsides:
- Along with updates to the Svelte tooling, Svelte typing has improved.
All exports made in a Svelte file are now visible to other files that
import them, and we no longer rebuild the Svelte files when TS files
are updated, as the Svelte files do no type checking themselves, and
are just a simple transpilation. Svelte-check now works on Windows again,
and there should be no errors when editing in VS Code after you've
built the project. The only downside seems to be that cmd+clicking
on a Svelte imports jumps to the .d.ts file instead of the original now;
presumably they'll fix that in a future plugin update.
- Each subfolder now has its own tsconfig.json, and tsc can be called
directly for testing purposes (but beware it will place build products
in the source tree): ts/node_modules/.bin/tsc -b ts
- We can drop the custom esbuild_toolchain, as it's included in the
latest rules_nodejs.
Other changes:
- "image_module_support" is moved into lib/, and imported with
<reference types=...>
- Images are now imported directly from their npm package; the
extra copy step has been removed.
Windows users may need to use "bazel clean" before building this,
due to old files lying around in the build folder.
The v3 scheduler will delay the final card from being shown twice in
a row, but the overdue case was being treated the same as the no-learning
case, leading to the message being hidden.
In order to split backend.proto into a more manageable size, the protobuf
handling needed to be updated. This took more time than I would have
liked, as each language handles protobuf differently:
- The Python Protobuf code ignores "package" directives, and relies
solely on how the files are laid out on disk. While it would have been
nice to keep the generated files in a private subpackage, Protobuf gets
confused if the files are located in a location that does not match
their original .proto layout, so the old approach of storing them in
_backend/ will not work. They now clutter up pylib/anki instead. I'm
rather annoyed by that, but alternatives seem to be having to add an extra
level to the Protobuf path, making the other languages suffer, or trying
to hack around the issue by munging sys.modules.
- Protobufjs fails to expose packages if they don't start with a capital
letter, despite the fact that lowercase packages are the norm in most
languages :-( This required a patch to fix.
- Rust was the easiest, as Prost is relatively straightforward compared
to Google's tools.
The Protobuf files are now stored in /proto/anki, with a separate package
for each file. I've split backend.proto into a few files as a test, but
the majority of that work is still to come.
The Python Protobuf building is a bit of a hack at the moment, hard-coding
"proto" as the top level folder, but it seems to get the job done for now.
Also changed the workspace name, as there seems to be a number of Bazel
repos moving away from the more awkward reverse DNS naming style.
I mourn the time lost trying to track this down :-(
https://github.com/protobufjs/protobuf.js/issues/1014
We can't patch the minified file in dist without essentially duplicating
it, so this change also switches from the external file to including
the src file as part of the bundle.
- prettier's formatting has changed, so files needed to be reformatted
- dart is spitting out deprecation warnings like:
254 │ 2: $spacer / 2,
│ ^^^^^^^^^^^
╵
bazel-out/darwin-fastbuild/bin/ts/sass/bootstrap/_variables.scss 254:6 @import
ts/sass/button_mixins.scss 2:9 @use
ts/components/ColorPicker.svelte 2:5 root stylesheet
DEPRECATION WARNING: Using / for division is deprecated and will be removed in Dart Sass 2.0.0.
Recommendation: math.div($grid-gutter-width, 2)
Most of our changes have been upstreamed, but the toolchain change
probably won't be merged soon, and a separate git repo will make it
easier to track upstream changes.
@hgiesel output_css=True will need to be changed to output_css="foo.css"
instead
Instead of generating a fluent.proto file with a giant enum, create
a .json file representing the translations that downstream consumers
can use for code generation.
This enables the generation of a separate method for each translation,
with a docstring that shows the actual text, and any required arguments
listed in the function signature.
The codebase is still using the old enum for now; updating it will need
to come in future commits, and the old enum will need to be kept
around, as add-ons are referencing it.
Other changes:
- move translation code into a separate crate
- store the translations on a per-file/module basis, which will allow
us to avoid sending 1000+ strings on each JS page load in the future
- drop the undocumented support for external .ftl files, that we weren't
using
- duplicate strings in translation files are now checked for at build
time
- fix i18n test failing when run outside Bazel
- drop slog dependency in i18n module
- The previous commits moved the majority of the remaining global css
into components; move the remaining @emotion/css references into
ticks.scss and the styling of the Graph.svelte. This is not as elegant
as the emotion solution, but builds a whole lot faster, and most of
our styling can be scoped to a component anyway.
- Leave the .html files in ts/ for now. AnkiMobile uses them, and
AnkiDroid likely will in the future too. In the long run we'll likely
move to loading the JS into an existing page instead of loading a
separate page, but at that point we can just exclude the .html file from
copy_files_into_group() without affecting other clients.
Closes#1074
To support images on that screen, we'll first need to adjust the base url
for each platform, or rewrite the local image URLs, as otherwise they
are resolved to _anki/pages/...
