This slims down SharedStyleContext, in preparation for a few things.
First, I would like to eventually move the stylist to the document in Servo, in
order for it to hold the StyleSheetSet.
Also, this gets rid of a fair amount of overhead while creating it in stylo.
Fixes bug 1363245.
I've chosen this approach mainly because there's no other good way to guarantee
the model is correct than holding the snapshots alive until a style refresh.
What I tried before this (storing them in a sort of "immutable element data") is
a pain, since we call into style from the frame constructor and other content
notifications, which makes keeping track of which snapshots should be cleared an
which shouldn't an insane task.
Ideally we'd have a single entry-point for style, but that's not the case right
now, and changing that requires pretty non-trivial changes to the frame
constructor.
MozReview-Commit-ID: FF1KWZv2iBM
Signed-off-by: Emilio Cobos Álvarez <emilio@crisal.io>
layout: Force reflow in the sequential fallback of block format context
When reflowing a block format context during the inorder traversal,
propagate restyle damage manually to its children since they were
already reflowed. Also, test the border box to see if it can fit into
floats according to CSS 2.1 § 9.5.
Improves reddit and yahoo.
---
<!-- 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 #__ (github issue number if applicable).
<!-- Either: -->
- [X] 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. -->
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/16458)
<!-- Reviewable:end -->
When reflowing a block format context during the inorder traversal,
propagate restyle damage manually to its children since they were
already reflowed.
Just use WebRender's ClipId directly. This will allow us to create and
use ReferenceFrames in the future, if we need to do that. It will also
make it easier to have Servo responsible for creating the root
scrolling area, which will allow removing some old hacks in the future.
Instead of passing down a complex clipping region to each item, used
shared clipping to handle overflow:hidden and CSS clips. In addition to
being more efficient, this should also fix quite a few issues related
to absolutely positioned elements.
One existing reftest is slightly modified to avoid tickling a quirk
with the way that WebRender rasterizes masks. We are working out how to
best express these combined masks with the API or need to. The change
does not affect the original subject of the reftest.
Fixes#13109.
Fixes#10151.
Fixes#7575.
Fixes#8074.
Fixes#8780.
I wanted to add an environmental variable to disable the style sharing
cache for gecko, but the current pattern involves lazy_static!, which
involves an atomic operation on lookup, which is a bit hot to do each
time we try to share styles. This makes that work happen once per
process.
While we're at it, we also eliminate the 'unknown' dom depth for the
bloom filter. Computing depth has negligible cost relative to the
amount of work we do setting up the bloom filter at a given depth.
Doing it once per traversal should be totally fine.
I originally separated the elimination of unknown dom depth from the
traversal changes, but I got bloom filter crashes on the intermediate
patch, presumably because I didn't properly fix the sequential traversal
for this case. Given that the final state is green, I just decided to
squash and move on.
Fix#16165.
Also, it turns out that the CSSFontFaceRule IDL specified in the
css-fonts spec is not web-compatible.
Instead browsers implement a .style attribute like in CSSStyleRule:
https://github.com/w3c/csswg-drafts/issues/825
This in turn requires preserving data about which descriptors
were set or not (distinguishing unset from set to a value that happens
to be the initial value),
so this commit also makes every field `Option<_>`.
We will set the flag when the root node of the traversal has
has-animation-only-dirty-descendants flag or has animation restyle hints.
Also we will use this flag to detect whether we need to trigger CSS transitions
or not.
This whole method was a large `if let` block:
```rust
if let Some(mut root_flow) = self.root_flow.clone() {
// ...
}
```
Move that `if let` to callers to make it clear that the method
doesn’t need to be called when `self.root_flow` is `None`.
By itself, this commit doesn’t change anything.
But it enables the next one.
Now that cascade() gets a Device, we can use the default computed values from
there to avoid propagating that state all over the place.
Signed-off-by: Emilio Cobos Álvarez <emilio@crisal.io>
At least until we support scrollbars properly, this size is going to be the
correct one. I've left a TODO to grab the proper one once we support it.
This allows to trivially test viewport units for now.
MozReview-Commit-ID: JdaZ6WlZ2C6
Signed-off-by: Emilio Cobos Álvarez <emilio@crisal.io>
In support of this goal, the layout thread collects information about
CSS images that are missing image data and hands it off to the script
thread after layout completes. The script thread stores a list of
nodes that will need to be reflowed after the associated network
request is complete. The script thread ensures that the nodes are
not GCed while a request is ongoing, which the layout thread is
incapable of guaranteeing.
The image cache's API has also been redesigned in support of this
work. No network requests are made by the new image cache, since it
does not possess the document-specific information necessary to
initiate them. Instead, there is now a single, synchronous
query operation that optionally reserves a slot when a cache
entry for a URL cannot be found. This reserved slot is then
the responsibility of the queryer to populate with the contents
of the network response for the URL once it is complete. Any
subsequent queries for the same URL will be informed that the
response is pending until that occurs.
The changes to layout also remove the synchronous image loading
code path, which means that reftests now test the same code
that non-test binaries execute. The decision to take a screenshot
now considers whether there are any outstanding image
requests for layout in order to avoid intermittent failures in
reftests that use CSS images.
Sometimes clippy gets outdated by months, and its current support setup
means that each Servo component need to opt into it by depending on
the plugins crate manually, and not all components do that.
This includes a WR update that improves pixel snapping. Specifically,
the non-transform shader now snaps to device pixels in world space,
rather than local space.
This makes 31 new tests pass, and adds 4 new failures.
Of the new failures:
- The three background size tests appear to be an OSMesa accuracy
issue. Near the bottom of the image, a small number of pixels
are different from the reference by a value of 1.0 / 255.0. I've
confirmed on 3 real GPUs and also on llvmpipe that the tests
match on those devices. This would be a good case for fuzzy test
support in WPT.
- The other failing test appears to have unreasonable accuracy
requirements as it also fails on Gecko and Chromium in the same
way that the test now fails on WR.
layout: Remove cached thread local context from LayoutContext, and use LayoutContext for assign_inline_sizes()
<!-- Please describe your changes on the following line: -->
According to https://github.com/servo/servo/pull/3069 the cached thread local context is introduced for green threads. Now green threads has gone, and the existence of cache force us to create a `LayoutContext`, an `AssignISizes` and an `AssignBSizes` for each flow during parallel layout, so the pull request tries to remove it. And it also switch `assign_inline_sizes()` to accept a `LayoutContext` parameter, as according to my current design we need to do full layout to some flex items for column flexbox during assign isize traversal.
Part of #14123.
---
<!-- 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
- [X] These changes do not require tests because refactoring
<!-- Pull requests that do not address these steps are welcome, but they will require additional verification as part of the review process. -->
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/servo/15417)
<!-- Reviewable:end -->
