This shaves a fair amount of time of stylist rebuild.
Most of the extra remaining overhead are memory allocations...
I'll try to get those sorted out next, though fixing that may overlap quite a
bit with the work Cameron is doing in bug 1382925.
Source-Repo: https://github.com/servo/servo
Source-Revision: b11b65837930fff7ec793e935af46d7edbeccda5
`stylo_tests` currently requires a separate version of the `style` crate, compiled with the `testing` feature, so a function testing the size of specified values can be accessed. With a few tweaks, we can make the information needed for the test available to the `stylo_tests` crate directly, eliminating the need for a separately-compiled `style` crate.
This doesn't matter much for Servo itself (it might make CI times slightly faster?), but Gecko automation/development would like to run `stylo_tests`, and not having to compile two versions of the `style` crate (or have a dead, test-only function hanging around in the `style` crate) would be a win.
---
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] There are tests for these changes
Source-Repo: https://github.com/servo/servo
Source-Revision: a6369149dc5344b2b80a12fca1c43cf99c94fdc9
We implement compute_distance for Angle to avoid returning Err(()) from it, and then rewrite compute_squared_distance of AnimatedFilterLIst to avoid using unwrap() and make it simpler.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix [Bug 1384014](https://bugzilla.mozilla.org/show_bug.cgi?id=1384014).
- [X] These changes do not require tests because there is a DevTools test for this already.
Source-Repo: https://github.com/servo/servo
Source-Revision: 3ab6b1cf8fa64cb31e169eece9e5e8224e195149
This relands #17701, with a fix for the test failures.
Source-Repo: https://github.com/servo/servo
Source-Revision: f49650ce96578c200c71b0a8fa50094764a3c76d
std::ptr::eq and Arc::ptr_eq are now usuable, and we can replace a
panic!() with abort().
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes do not require tests because they are trivial replacements.
Source-Repo: https://github.com/servo/servo
Source-Revision: bb9391ba2cacc2976d546c42cb339fa3e7cbf35c
This is a rebase of #17325 with `[replace]` entries removed, a bunch more dependencies updated, and some more compile fixes. Original work by @Eijebong, thanks a lot!
Source-Repo: https://github.com/servo/servo
Source-Revision: 66c130d55aa0d7af1104c00e93a5bf950f23a383
This file has become quite bloated lately. This commit deletes that file in
favor of a set of submodules.
The only noticeable change apart from code move, is converting deep_clone_foo
methods into a trait.
It also unifies logic related to different style rules in the same place.
There's some missing work, specially related to font-face and counter-style, but
I think this is worth landing in the meantime.
Source-Repo: https://github.com/servo/servo
Source-Revision: c260c6ede0988f331011f33485964775e1f3a7f9
Depends on https://github.com/servo/unicode-bidi/pull/27 , which
upgrades `unicode-bidi` crate to `0.3.0`.
Summary of changes:
* Use `unicode_bidi::Level` (instead of `u8`) in all relevant places and
replace magic computations with (inline) method calls to Level API.
* Doing so required adding `unicode-bidi` crate dependency to two more
components here: `style` and `gfx`. IMHO, totally makes sense, as
replaces local integer manipulations/checks with well-tested ones
already available in a common dependency.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [ ] These changes fix #__ (github issue number if applicable). [N/A]
<!-- Either: -->
- [ ] There are tests for these changes OR
- [X] These changes do not require tests because `unicode-bidi` has its own tests and there's no logic change in this diff.
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Source-Repo: https://github.com/servo/servo
Source-Revision: edd6c2cecb2c245f9f8bcab04ff8e57f1c5d1333
This is just a code health change. I want to move it away to keep matching.rs as
simple as possible.
Source-Repo: https://github.com/servo/servo
Source-Revision: 03fbea4ec8af3ef6c5c621974b3c52cd755ed287
This enum type used to contain the result of parsing one CSS source declaration (`name: value;`) and expanding shorthands. Enum types are as big as the biggest of their variant (plus discriminant), which was quite big because some shorthands expand to many longhand properties. This type was returned through many functions and methods, wrapped and rewrapped in `Result` with different error types. This presumably caused significant `memmove` traffic.
Instead, we now allocate an `ArrayVec` on the stack and pass `&mut` references to it for various functions to push into it. This type is also very big, but we never move it.
We still use an intermediate data structure because we sometimes decide after shorthand expansion that a declaration is invalid after all and that we’re gonna drop it. Only later do we push to a `PropertyDeclarationBlock`, with an entire `ArrayVec` or nothing.
In future work we can try to avoid a large stack-allocated array, and instead writing directly to the heap allocation of the `Vec` inside `PropertyDeclarationBlock`. However this is tricky: we need to preserve this "all or nothing" aspect of parsing one source declaration, and at the same time we want to make it as little error-prone as possible for the various call sites. `PropertyDeclarationBlock` curently does property deduplication incrementally: as each `PropertyDeclaration` is pushed, we check if an existing declaration of the same property exists and if so overwrite it. To get rid of the stack allocated array we’d need to somehow deduplicate separately after pushing multiple `PropertyDeclaration`.
Source-Repo: https://github.com/servo/servo
Source-Revision: 60682cf81fe19a82c73dd98ba4c1eebc1dbbfcac
This cuts in almost half the time to run:
```
touch components/style/lib.rs
./mach test-stylo
```
Source-Repo: https://github.com/servo/servo
Source-Revision: a01ab9ad3462bb5f30f5ea7903abd6deb9beb364
This will allow reusing it from text styles, which we need for some corner
cases, like text-align: -moz-center and similar stuff.
Source-Repo: https://github.com/servo/servo
Source-Revision: 09f0ff74811405eb95200ee78d936b4378de65fb
This is needed for both bug 1357461 and bug 1273303, where I plan to add smarter
invalidations than what we have now.
Also, it's cleaner.
Ideally I'll move this onto stylist, though that may require extra work to make
it work fine for Servo, so for now let's just do the obvious thing.
Source-Repo: https://github.com/servo/servo
Source-Revision: 5274991545fedd9b7188c40abffb61b4067daa8a
It seems mod attr is not used for geckolib at all, and that is the only place where servo_url is still referenced for geckolib, so we can just remove it.
Source-Repo: https://github.com/servo/servo
Source-Revision: 12a4cc875d44a1a516412dc4c6044e897c0c6e47
To be aligned with the implementation from Gecko side, we parse
font-language-override as Normal keyword or String. Then, we compute
and store it as a u32. So, as to the stylo glue, we can just pass the
u32 to Gecko.
The extra crate, byteorder, is used to simplify the computing and
serialization.
Since we now implement font-language-override for Gecko, we can remove
the additional branches for font-language-override in font shorthand.
ref: Gecko [Bug 1347821](https://bugzilla.mozilla.org/show_bug.cgi?id=1347821)
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [ ] These changes fix #__ (github issue number if applicable).
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because _____
<!-- Also, please make sure that "Allow edits from maintainers" checkbox is checked, so that we can help you if you get stuck somewhere along the way.-->
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Source-Repo: https://github.com/servo/servo
Source-Revision: af243d5decde5663e2f6558eaa78c9022536c289
<!-- Please describe your changes on the following line: -->
Before this PR, every object reflected in CSSOM is in `Arc<RwLock<_>>` to enable safe (synchronized) mutable aliasing. Acquiring all these locks has significant cost during selector matching:
* https://bugzilla.mozilla.org/show_bug.cgi?id=1311469
* https://bugzilla.mozilla.org/show_bug.cgi?id=1335941
* https://bugzilla.mozilla.org/show_bug.cgi?id=1339703
This PR introduce a mechanism to protect many objects with the same `RwLock` that only needs to be acquired once.
In Stylo, there is one such lock per process (in a `lazy_static`), used for everything.
I non-Stylo Servo, I originally intended to have one such lock per document (for author-origin stylesheets, and one per process for user-agent and user sytlesheets since they’re shared across documents, and never mutated anyway). However I failed to have the same document-specific (or pipeline-specific) `Arc` reachable from both `Document` nodes and `LayoutThread`. Recursively following callers lead me to include this `Arc` in `UnprivilegedPipelineContent`, but that needs to be serializable. So there is a second process-wide lock.
This was previously #15998, closed accidentally.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [ ] These changes fix #__ (github issue number if applicable).
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because _____
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Source-Repo: https://github.com/servo/servo
Source-Revision: bb54f0a429de0e8b8861f8071b6cf82f73622664
<!-- Please describe your changes on the following line: -->
https://github.com/servo/rust-cssparser/pull/123
In particular, `match_ignore_ascii_case` now supports the full `match` syntax.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [ ] These changes fix #__ (github issue number if applicable).
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because _____
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Source-Repo: https://github.com/servo/servo
Source-Revision: fbfcfc2dbe838ef011f14a863003a575078f455f
<!-- Please describe your changes on the following line: -->
<s>Depends on https://github.com/servo/rust-cssparser/pull/122.</s>
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [ ] These changes fix #__ (github issue number if applicable).
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because _____
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Source-Repo: https://github.com/servo/servo
Source-Revision: 94e563e4d9292d7b19ce061e070cda358e822172
By having a single thread pool, rather than one per document, we use less memory. This addresses https://bugzilla.mozilla.org/show_bug.cgi?id=1324250.
This may be obvious to an experienced Rust programmer, but I went with raw pointers because trying to use `Option` global variables resulted in complaints about turning on feature flags in nightly Rust. Since this is for stylo, nightly features are not appropriate here.
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [ ] These changes fix #__ (github issue number if applicable).
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because _____
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Source-Repo: https://github.com/servo/servo
Source-Revision: 56a99577b31a942e340624f97377980b0e612088
<!-- Please describe your changes on the following line: -->
I have been working on issue number #15244. I have substituted quickersort with pdqsort in the style component. The changes I made affect: Cargo.toml, stylist.rs and lib.rs (all of them are in /components/style).
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [X] `./mach build -d` does not report any errors
- [X] `./mach test-tidy` does not report any errors
- [X] These changes fix#15244 (github issue number if applicable).
<!-- Either: -->
- [ ] There are tests for these changes OR
- [X] These changes do not require tests because both library expose the same functionality, they differ in permance.
Source-Repo: https://github.com/servo/servo
Source-Revision: d4ee8a3599a57078735766640b31df31c67d8201
<!-- Please describe your changes on the following line: -->
---
<!-- Thank you for contributing to Servo! Please replace each `[ ]` by `[X]` when the step is complete, and replace `__` with appropriate data: -->
- [x] `./mach build -d` does not report any errors
- [x] `./mach test-tidy` does not report any errors
- [ ] These changes fix #__ (github issue number if applicable).
<!-- Either: -->
- [ ] There are tests for these changes OR
- [ ] These changes do not require tests because _____
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
Source-Repo: https://github.com/servo/servo
Source-Revision: df895cb80c460594f220b6cf28aa5869464d05ec
The idea here is having (for convenience) different `Expression` and `Device` representations for Gecko.
Both will be implemented in rust, but will use `nsMediaFeatures` as a source of media features instead of the Servo bits.
Does it sound good?
r? @heycam or @Manishearth
Source-Repo: https://github.com/servo/servo
Source-Revision: fdddeb1c9229b6f9bc202cb2df1a3701eec47e05
