Automatic update from web-platform-tests
[layout] Fix null-dref crash with removed layout object.
As above - we were trying to access a layout object which was
cleared from the fragment.
Fixed: 40286186
Change-Id: Ibae87e40759f1e9ebe39062c4ba6d8752b9a8125
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5760561
Reviewed-by: David Grogan <[email protected]>
Commit-Queue: Ian Kilpatrick <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1337108}
--
wpt-commits: a7eb020c78e4ef449ee89e67fe7096af84e86409
wpt-pr: 47446
Automatic update from web-platform-tests
Remove NonStandardAppearanceValuesHighUsage, shipped in 124
The deprecation was completed and fully disabled for M124 users [1].
This CL removes all of the code supporting the deprecated high usage
appearance values.
Note that part of blink code still uses the ControlPart values to
render some widgets (e.g. inner-spin-button).
This patch is only to remove the support for these as CSS values.
We also replaced the Virtual test suite
"deprecated-css-appearance-values-enabled" by "slider-vertical-enabled",
where only feature NonStandardAppearanceValueSliderVertical is enabled.
[1] https://chromium-review.googlesource.com/c/chromium/src/+/5341111
Change-Id: I277302655e7b72ce9b5b7eb644d57046fd44b7f5
Bug: 40610694
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5757984
Commit-Queue: Di Zhang <[email protected]>
Reviewed-by: Mason Freed <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1336897}
--
wpt-commits: 1eaddc37a55977200ae3f983eafc1cfcb121235e
wpt-pr: 47436
Automatic update from web-platform-tests
Rename position-try-options to position-try-fallbacks
This has 3 LGTMs to ship:
https://groups.google.com/a/chromium.org/g/blink-dev/c/kJwhHYXupWk/m/qYfaEe98CAAJ
Note that this CL also adds a virtual test suite that re-enables
the `position-try-options` property. I wanted to include that in
the same CL, because I plan to try to merge this CL back to M128,
and I'd like the test coverage to follow.
Fixed: 349600667
Change-Id: Ibd6415c1da80a1e1ca34874fece9fcf360d40668
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5739360
Reviewed-by: Mike Taylor <[email protected]>
Commit-Queue: Mike Taylor <[email protected]>
Auto-Submit: Mason Freed <[email protected]>
Reviewed-by: Rune Lillesveen <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1334247}
--
wpt-commits: 3bf9755fce4f412f145bb2d7f9c0579fc4f7cc92
wpt-pr: 47330
Automatic update from web-platform-tests
Move tests to match the spec's move
See https://github.com/w3c/csswg-drafts/pull/10447
--
wpt-commits: c6273f7c688e8e62356f2a0a6e388108d58042ef
wpt-pr: 46928
Automatic update from web-platform-tests
[@container] Clear skipped style recalc for all descendants
The existing code assumed we would always resume style recalc from the
skipped container's layout, but that is not always the case.
Take the following scenario which could happen with nested containers:
1. Style recalc is not skipped for an outer container because it is not
marked for layout.
2. Style recalc is skipped for an inner container because it is marked
for layout.
3. Layout changes the width of the outer container because the outer
container is auto-sized, and changes size because an ancestor box
changes its size.
4. The @container evaluation changes for the outer container, which
means we enter UpdateStyleAndLayoutTreeForContainer() for the outer
container.
5. We reach recalc for the inner container, whose subtree was skipped
for style recalc, but we did not clear the skipping or retrieve the
stored style recalc change.
In this CL we always check, clear, and return the StyleRecalcChange if
style recalc was skipped.
Bug: 346264227
Change-Id: I3dc997850068a6751333ed588adf2f1f0e87518f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5630638
Reviewed-by: Anders Hartvoll Ruud <[email protected]>
Commit-Queue: Rune Lillesveen <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1317345}
--
wpt-commits: 1ec7cf9a1bf651899999bab7f99e09d48416d708
wpt-pr: 46843
Automatic update from web-platform-tests
[@container] Container lookup in flat tree
The spec changed from looking up shadow-including ancestors to flat tree
ancestors for finding the container being queried by a container query.
Bug: 340876720
Change-Id: Id069f09c311fd49256c73bcf6d4b9b420ebe6c7c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5625932
Commit-Queue: Rune Lillesveen <[email protected]>
Reviewed-by: Anders Hartvoll Ruud <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1315123}
--
wpt-commits: 5f7fee3064a32d5d175545fc8f4019a39ea12194
wpt-pr: 46747
Automatic update from web-platform-tests
Update scroll to text fragment test mapping (#46249)
The feature was renamed in web-features:
https://github.com/web-platform-dx/web-features/pull/1007
Additionally include some tests of the interaction with other features.
This results in some tests being tagged with multiple features, which
could potentially be used to filter out tests depending on the
implementation status of the different features.
--
wpt-commits: dab963f555e6136cf4a51e4f162937996f3ce732
wpt-pr: 46249
Automatic update from web-platform-tests
[layout] Fix synthesis of inline baselines.
Previously for <button>, <input type=button> (and friends), and
<input type=range> we'd explicitly produce a baseline at appropriate
"end" edge.
This wasn't strictly correct instead we should be synthesizing them
within LogicalBoxFragment::BaselineMetrics.
There are three different ways to synthesize these baseline, off the
margin-box (the default), content-box (buttons), and border-box (range
and checkbox/radio with effective appearance).
Instead of checking all these conditions, this patch introduces an
additional field to the computed style to tell BaselineMetrics which
edge the synthesize off.
This patch changes two WPT tests which I believe were incorrect, see:
https://github.com/web-platform-tests/wpt/issues/45889
The TL;DR is that:
<button></button> <button style="contain:layout"></button>
Currently have different baselines when they shouldn't.
The image rebaselines are ensuring that we synthesize the baseline
for range controls consistently. Previously we'd synthesize them off
the border-box if they were in the same writing-mode, and off the
margin-box if in a different writing-mode.
Bug: 40937312
Change-Id: I28eedfbce58ead9d5e3119ae3b164c559fb9267c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5479489
Reviewed-by: Morten Stenshorne <[email protected]>
Commit-Queue: Ian Kilpatrick <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1293898}
--
wpt-commits: b312464e2203190809de7c34055a51b7643ba6c2
wpt-pr: 45936
Automatic update from web-platform-tests
Always initialize PaintInfo::descendant_painting_blocked_
`PaintInfo::descendant_painting_blocked_` is a cache of
`LayoutObject::ChildPaintBlockedByDisplayLock()`, but wasn't being
initialized in `BoxFragmentPainter::PaintTextClipMask`, which led to
painting locked content. This patch prevents this class of bugs by
forcing `descendant_painting_blocked_` to be initialized in the ctor.
Bug: 333952115
Change-Id: Ibea822ba510e1c854d63c5f6dd2b025535498abc
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5472456
Reviewed-by: Vladimir Levin <[email protected]>
Commit-Queue: Philip Rogers <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1291343}
--
wpt-commits: ec57fe8800ad35b9648561988ebeb049f5bb58df
wpt-pr: 45835
Automatic update from web-platform-tests
Support computing light-dark() for registered color syntax
ComputeRegisteredPropertyValue() did not handle light/dark pairs.
Added some more complex computed style tests for registered color
syntax. color-mix() is not supported either. To be fixed in a separate
CL.
Bug: 335383222
Change-Id: Ic7c4dddeb141b71e16aa86fee4df8ef3c5b5bd48
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5463142
Commit-Queue: Rune Lillesveen <[email protected]>
Reviewed-by: Anders Hartvoll Ruud <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1289903}
--
wpt-commits: b7ecf93438c84775b6e7bfc4a53a7481fecabf6e
wpt-pr: 45791
Automatic update from web-platform-tests
Tests for an explicit `initial` and “space” style queries (#45759)
--
wpt-commits: 13789bb2123e5a84de327f8dd736629ae8243ecd
wpt-pr: 45759
Automatic update from web-platform-tests
Paint selection with other highlight parts
When using the highlight overlay painting path, selection was
being painted as a separate step after the other highlight parts.
This resulted in double painting of non-selected text after a
recent CL, and also prevents selection from inheriting non-specified
colors from layers below.
This CL moves selection painting into the same loop as other highlight
parts, removing the double painting and making it possible, in the
next CL, to fix the color inheritance problem.
As a necessary substep:
* Add a LineRelativeRect::Unite method
* Add methods to adjust a line relative rect to match ink overflow
* Add unit testing for LineRelativeRect
* Add a cc file for LineRelativeRect.
And remove some of the fuzzy matching from a previous CL.
Bug: 40264572
Change-Id: Ic9a578895f363c85b0c2d2bb4e67dcb5b4159f5f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5447735
Reviewed-by: Fredrik Söderquist <[email protected]>
Commit-Queue: Stephen Chenney <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1288710}
--
wpt-commits: 22fe04d3f6f32ae698aa74ae35fb2db189d10cb4
wpt-pr: 45716
Automatic update from web-platform-tests
Fix rendering of font ligatures with highlights
Highlights for a range that falls within a font ligature should
behave as selection behaves, and split the ligature. To achieve
this, apply the selection painting approach to highlight text.
Use clipping to the highlight rect and paint text that
expands to include all the ligature characters before and after
the range.
To then render the originating text correctly outside the
highlighted range, we must also move the originating text painting
into the same loop as the highlight parts.
All this is in preparation for moving selection into the same
loop.
Bug: 40264572
Change-Id: I70e2c333cdb76f1a73612f186254b668324048e2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5430353
Auto-Submit: Stephen Chenney <[email protected]>
Reviewed-by: Fredrik Söderquist <[email protected]>
Commit-Queue: Stephen Chenney <[email protected]>
Commit-Queue: Fredrik Söderquist <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1285776}
--
wpt-commits: 1d522695b62ff465480400962a1dbc1ab923b41c
wpt-pr: 45596
Automatic update from web-platform-tests
Fix style containment scope removal
When removal is coming from Node::WillBeRemoved, we don't need to
reattach anything, just clear the scope and remove it from scopes,
unlike reattaching as in case of style update.
Fixed: 329231572
Change-Id: I70fb8d9b876ee04bac52f022430ca8aa65316239
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5380324
Reviewed-by: Rune Lillesveen <[email protected]>
Commit-Queue: Daniil Sakhapov <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1275594}
--
wpt-commits: fa118071ec249c4f2a034b9977bddba983aeb4ad
wpt-pr: 45199
Automatic update from web-platform-tests
Fuzz Fluent failing wpt tests.
This CL adds a temporary fuzz acceptance to some wpt web tests. The
fuzziness is added to make web tests with fluent scrollbars pass when
there are differences between the baseline and the ref html that cause
some scrollbars to be composited and others to not be, which causes a
1 shade of gray difference between some anti-alias pixels in the
scrollbar's thumb.
These tests are failing in wpt.fyi for Edge, since Edge has Fluent
scrollbars enabled by default.
This issue will be fixed by
https://issues.chromium.org/issues/40778657 , and when that work is
finished I will revert this CL.
Bug: 326116366, 329199184
Change-Id: Ib8e26c1bdcdfa3f78d7eab198dd62b290d79304d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5362506
Reviewed-by: Olga Gerchikov <[email protected]>
Commit-Queue: Gaston Rodriguez <[email protected]>
Reviewed-by: Yaroslav Shalivskyy <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1271612}
--
wpt-commits: 7557e87f6686cb37c27cb789d49d91c7630bdad5
wpt-pr: 45045
This is kind of a wallpaper, but I think it's sound, read below.
The reason this is asserting is that MathML in particular has some odd
frame re-parenting going on (see
nsCSSFrameConstructor::FlushAccumulatedBlock).
That means that our assert in DepthOrderedFrameList is not really
holding depending on when you add or remove the frame from the list.
However this is not a problem in practice, at least for container
queries, because anonymous frames that can wrap mathML cannot be
containers themselves.
I need to figure out what to do with that assert, but given it's useful
for the other users of DepthOrderedFrameList, I think we should take
this for now.
Differential Revision: https://phabricator.services.mozilla.com/D200634
Automatic update from web-platform-tests
DL: Update display lock condition to use EDisplay::kRubyText.
This patch updates the condition for when to ignore a display lock
to use a ruby-text display mode, which is consistent with the ruby
check in IsStackingContext.
Fixed: 1515531
Change-Id: Ide6b1a3e76bf414e6dcbd1833c96735051cb2902
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5213689
Reviewed-by: Xianzhu Wang <[email protected]>
Commit-Queue: Vladimir Levin <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1249506}
--
wpt-commits: 3c85ffc526170187bb331c1716dd7fdaf11f11b5
wpt-pr: 44084
Automatic update from web-platform-tests
Fix content-visibility-095.html expectation
This test seems to be working, except that the reference image doesn't
apply `contain: content` to the table corresponding to table.hidden,
which has content-visibility:hidden.
Bug: 1510865
Change-Id: I3beb913ad7910f983aec15c262d06e282006e063
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5195268
Auto-Submit: Mason Freed <[email protected]>
Reviewed-by: Ian Kilpatrick <[email protected]>
Commit-Queue: Ian Kilpatrick <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1246889}
--
wpt-commits: 5049a31d2a7eebd7bca08317e56664021a8bd36c
wpt-pr: 43983
Automatic update from web-platform-tests
Fix Highlight Pseudos using container-relative units
Add support for creating a new style when the highlight uses
container-relative units and has a different container to that
of its parent. And also fix a problem where highlight styles
are not created inside container queries for the container itself.
Add testing of highlights using container queries.
Fixes a typo too.
Units depending on block/inline direction still do
not work for vertical writing modes. That's for another CL.
Bug: 1468306
Change-Id: I781a35810133f3591c7a3587ff5d13df0dfac9ee
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5149171
Reviewed-by: Rune Lillesveen <[email protected]>
Commit-Queue: Stephen Chenney <[email protected]>
Cr-Commit-Position: refs/heads/main@{#1245980}
--
wpt-commits: 217ab9f1480ea3ebce180e7cf9f787246691bfe7
wpt-pr: 43848
These tests verify that floats and margin-collapsing are properly
handled when paint/layout containment changes dynamically.
Differential Revision: https://phabricator.services.mozilla.com/D197885