Automatic update from web-platform-tests
[FlexNG] Fix great-grandchild relative-position offsets.
In order to determine the correct size to resolve %-based relative
top offsets against we need to determine if the available-size given
to a flex-item is definite.
Previously this worked as there was a bit on the LayoutBox which
captured this information. This patch for NG flex-items allows for
looking at the previous constraint space while outside of layout.
Bug: 1108508
Change-Id: I3104b7afd4d96b707a8110eb2ed5d59d17129580
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2327838
Commit-Queue: Ian Kilpatrick <ikilpatrick@chromium.org>
Reviewed-by: David Grogan <dgrogan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#793028}
--
wpt-commits: fd5a851b2e6bd60ff49060f1b42197c976273728
wpt-pr: 24805
Automatic update from web-platform-tests
[FlexNG] Flex items can't be relayout roots, even with `contain`
When a flex item is a relayout root, it gets neither the flexed nor
stretched dimensions from the parent flexbox. Though this problem seems
to manifest only when the flex item is a legacy layout node.
This change causes a 5% regression in the change-text-css-contain perf
test, even though it doesn't use flex, probably just because we test for
flex-item before deciding that a node with `contain: layout size` can be
a layout root.
Bug: 1108675
Change-Id: Iec0d35557ebfbf3136bd66fe986ba2456e0d92ab
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2322074
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Commit-Queue: David Grogan <dgrogan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#792833}
--
wpt-commits: 3b29525440ec57478a885a7d34f7475a4eb38854
wpt-pr: 24786
Automatic update from web-platform-tests
webrtc wpt: fix split.https.html firefox compatibility
The test does not work as intended in Firefox as described in
https://bugzilla.mozilla.org/show_bug.cgi?id=1643050
since Firefox does not use the same fingerprint/certificate for different
peerconnections on the same page and puts the fingerprint into the SDP sessionpart.
With the change the test connects and gets video on the other end but still
fails in Firefox since it relies on connectionState which is not implemented.
BUG=None
Change-Id: Id6529bcb32f37e657d6f79c8d94a4b69c4d15db3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2303391
Reviewed-by: Harald Alvestrand <hta@chromium.org>
Commit-Queue: Philipp Hancke <philipp.hancke@googlemail.com>
Cr-Commit-Position: refs/heads/master@{#792646}
--
wpt-commits: 21d0ad202f409a6e955ae3d000490db943bfaff3
wpt-pr: 24645
Automatic update from web-platform-tests
Added 9 new active-selection-* tests and their references (#24788)
--
wpt-commits: 9a60de15eda095faa363e76eec7cdc04fac36704
wpt-pr: 24788
Automatic update from web-platform-tests
Fix layout shift tracking with 2d transforms
This fixes the following issues:
1. crbug.com/1109053 we reported transform changes as layout shifts.
According to https://github.com/WICG/layout-instability#transform-changes
we should ignore transform changes.
This is fixed by just accumulating offsets of PaintOffsetTranslations
for the additional offset for layout shift
(previous offset_to_2d_translation_root,
now additional_offset_to_layout_shift_root)
Test: external/wpt/layout-instability/transform-change.html
2. Movement of transformed element itself was ignored. This is fixed by
resetting the additional offset for layout shift after updating
paint properties for child instead of during updating the properties
because the offset applies to the transformed element itself.
Test: external/wpt/layout-instability/move-transformed.html
3. Stale old additional offset for layout shift (previously baked in
FragmentData::VisualRectTo2DTranslationRoot). We don't always update
the offset when it changes because it's neither a paint property nor
a paint invalidation data, thus the old data may be stale. To avoid
that, now save VisualRectForLayoutShiftTracking() which is in the
local transform space and is always updated, and recover the old
additional offset from the old paint properties.
Bug: 1109053,1104794
Change-Id: I36a3dec0695f9674251e7f5ce85ae2411b135eda
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2316783
Reviewed-by: Chris Harrelson <chrishtr@chromium.org>
Reviewed-by: Steve Kobes <skobes@chromium.org>
Commit-Queue: Xianzhu Wang <wangxianzhu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#792312}
--
wpt-commits: 321d9573a97057a9b25bf95d9e1b449f541537f6
wpt-pr: 24757
Automatic update from web-platform-tests
Specify channel for stable Safari runs on Azure Pipelines (#24785)
This should never have worked (yay), as the default for channel is `nightly`, but used to
do so before some recent refactoring. Properly specifying `--channel=stable` for Safari
should fix this.
See https://github.com/web-platform-tests/wpt/issues/24784
--
wpt-commits: 7c7fe3fed7aef60746613acca3e896380594d90c
wpt-pr: 24785
Automatic update from web-platform-tests
[wpt] Remove a redundant channel arg (#24783)
from Chrome.find_nightly_binary and fix call sites.
Add a test, too, which helped to identify and fix a bug on macOS.
--
wpt-commits: 9a1bdfa6d5f1aa8dbb5094af1fad1dd8c5e681af
wpt-pr: 24783
Automatic update from web-platform-tests
Python 3: Drop use of flush on ResponseWriter in a couple of files (#24657)
This patch only drops the use of response.write.flush() in a couple of
files as they directly affect test results in Python 3.
Detailed discussions on dropping use of flush on ResponseWriter can be
found at https://github.com/web-platform-tests/wpt/pull/24617
(Also see #24719 )
--
wpt-commits: a5583f1f9f1eb9054b1d587abe54d05d7036ea28
wpt-pr: 24657
Automatic update from web-platform-tests
Python 3: port tests in cookie-store (#24531)
Co-authored-by: Robert Ma <robertma@chromium.og>
--
wpt-commits: 79d185a6f5cb386314e8052c874f4492bb1e04d8
wpt-pr: 24531
Automatic update from web-platform-tests
[COOP] access reporting. Remove excessive test cases.
The [window.idl] defines CrossOrigin indexed and named 'getter', but it
doesn't define their 'setter' counterpart.
As a result, we should remove test for them. They were added in excess.
[window.idl]: third_party/blink/renderer/core/frame/window.idl
Bug: chromium:1090273
Change-Id: I31fb77c65b82fef3d6f44297bfa63a331a9e1b02
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2315694
Reviewed-by: Yuki Shiino <yukishiino@chromium.org>
Reviewed-by: Camille Lamy <clamy@chromium.org>
Auto-Submit: Arthur Sonzogni <arthursonzogni@chromium.org>
Commit-Queue: Arthur Sonzogni <arthursonzogni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#792214}
--
wpt-commits: 49aff0cd973d56736d0d9462b4f4d099ec38bb55
wpt-pr: 24722
Automatic update from web-platform-tests
Move WebAssembly toStringTag test and run it in js shells.
--
wpt-commits: 0dae4d670e27cf3f8ee42bf31292c70d0ef85db5
wpt-pr: 24778
Automatic update from web-platform-tests
[wptpr.live] Make sure to clean up old refs/prs-open/{pr} (#24768)
This expands the list of PRs that pr_preview.py looks at to include
those that have existing refs. That means that old PRs outside the
window will always have their deployments deleted, rather than being
kept forever.
Fixes https://github.com/web-platform-tests/wpt.live/issues/36
--
wpt-commits: d409a92831a5e9823d123489e606a308d4272648
wpt-pr: 24768
Automatic update from web-platform-tests
[FlexNG] Fix flex-item grid layout explosion.
When we had a column flex-item, which was also a grid, where the percent
template added up to something >100% on each layout the grid would grow
larger.
This was caused by some logic to fix relative-positioned children in:
https://chromium-review.googlesource.com/c/chromium/src/+/2093273
However when this value was read within layout, it could refer to the
previous layout. If numerous layouts occurred sequentially, this would
cause the layouts to grow unbounded.
This patch disables this path when read during layout for a flex-item
(in NG). During layout the flex-item should have a BoxLayoutExtraInput
set.
Bug: 1108928
Change-Id: I4d0d41755c9d33eb2b4c217ee6bfdb1acd48cbbf
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2320418
Reviewed-by: David Grogan <dgrogan@chromium.org>
Commit-Queue: David Grogan <dgrogan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#792193}
--
wpt-commits: f153a0513f12e53717c45b5bf1f1aaa2e3d02f87
wpt-pr: 24773
Automatic update from web-platform-tests
Revert "[COOP] Access reporting. Fix the python server."
This reverts commit 9c96261a884c9ebd9daff66b7e41398d65c9d11c.
Reason for revert: This appears to have broken external/wpt/html/cross-origin-opener-policy/reporting/access-reporting/openee-accessed_openee-coop-ro.https.html
BUG: 1110254
Original change's description:
> [COOP] Access reporting. Fix the python server.
>
> A python server is used has a communication channel in between documents
> from different browsing context group and potentially cross-origin.
>
> There are two methods:
> - one for writing a message.
> - one for reading the message.
>
> I initially supposed we won't need to store more than one message in the
> stash. It turns out there are tests doing two write in a row without any
> read in the middle. It was throwing an exception when this
> happened.
>
> This patch makes the server to maintain a queue of messages instead. It
> doesn't throw errors anymore.
>
> Bug: chromium:1109194
> Change-Id: I25f8dc2c2a032781a4527a0994e94e46f85e12c9
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2317865
> Reviewed-by: Pâris Meuleman <pmeuleman@chromium.org>
> Commit-Queue: Arthur Sonzogni <arthursonzogni@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#791901}
TBR=clamy@chromium.org,arthursonzogni@chromium.org,pmeuleman@chromium.org
Change-Id: I87ba2caa07f80f42aa549929ade392485b968d6f
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:1109194
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2320572
Reviewed-by: Tim Schumann <tschumann@chromium.org>
Commit-Queue: Tim Schumann <tschumann@chromium.org>
Cr-Commit-Position: refs/heads/master@{#792184}
--
wpt-commits: 8a0eb4582d5f63590ad50cc8ccbd7633ffd17086
wpt-pr: 24777
Automatic update from web-platform-tests
SetNeedsPaint on tables when collapsed borders change
Tables always paint collapsed borders, even if the rows are seprately
composited. While we don't need to invalidate paint for the table
(because the row is the DisplayItemClient for the borders), we do need
to cause repaint for its painting layer, because the table paints those
borders.
See also (*), which removed the code that previously had the same
effect. There was previously no test for this code.
(*) https://chromium-review.googlesource.com/c/chromium/src/+/2317848
Change-Id: I269101d3bf28445c0ec3d1f6bcad1a23cf3e46d4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2321234
Commit-Queue: Chris Harrelson <chrishtr@chromium.org>
Reviewed-by: Xianzhu Wang <wangxianzhu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#792138}
--
wpt-commits: ff1fd0c93e31a8525e4bfbf1f19edc0736e0a792
wpt-pr: 24771
Automatic update from web-platform-tests
Implement ScrollTimeline.scrollOffsets API
This change implements initializing ScrollTimeline with scrollOffsets
and reading back scrollOffsets.
Bug: 1094014
Change-Id: I9d39bcc2644cee1288d59a29119d9ac1d91b5080
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2264926
Reviewed-by: Robert Flack <flackr@chromium.org>
Commit-Queue: Olga Gerchikov <gerchiko@microsoft.com>
Cr-Commit-Position: refs/heads/master@{#792106}
--
wpt-commits: fb149eacad8873f8b1a6215063d1f6092b5f9821
wpt-pr: 24341
Automatic update from web-platform-tests
Have CookieStore reject cookies with no name and no value.
The CookieStore spec recently changed to disallow setting cookies
using this API:
https://github.com/WICG/cookie-store/pull/150
This CL updates the implementation of CookieStore in Blink to
perform this additional check and updates the web platform
tests to reflect the new changes to the spec.
Bug: 1108964
Change-Id: I1bc60332551308498509c0e8bbc6823bd8375ab8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2316622
Reviewed-by: Ayu Ishii <ayui@chromium.org>
Commit-Queue: Dylan Cutler <dylancutler@google.com>
Cr-Commit-Position: refs/heads/master@{#791963}
--
wpt-commits: de61dd6ef3fe74554b638d741dc2b68bc44ad65f
wpt-pr: 24730
Automatic update from web-platform-tests
Fix test of non-animatable properties.
Follow-up to https://chromium-review.googlesource.com/c/chromium/src/+/2303353
addressing post-submission comments.
Add missed test case and fix list sorting.
Bug: 1105144
Change-Id: I307065b3a8f4915a8b400c328ef2998fa8305e7c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2314296
Reviewed-by: Oriol Brufau <obrufau@igalia.com>
Reviewed-by: Kevin Ellis <kevers@chromium.org>
Commit-Queue: George Steel <gtsteel@chromium.org>
Cr-Commit-Position: refs/heads/master@{#791933}
--
wpt-commits: 08846fbf9a3c39da2427d06e742a285c3400ee95
wpt-pr: 24700
Automatic update from web-platform-tests
[wpt] Fix the installation of ChromeDriver stable (#24766)
This was accidentally regressed in #24702 due to lack of comments or
tests, both of which are also added in this change.
Fixes#24753.
--
wpt-commits: 93f60af091e034be401e3915c3d699d495e7a5ea
wpt-pr: 24766
Automatic update from web-platform-tests
[COOP] Access reporting. Fix the python server.
A python server is used has a communication channel in between documents
from different browsing context group and potentially cross-origin.
There are two methods:
- one for writing a message.
- one for reading the message.
I initially supposed we won't need to store more than one message in the
stash. It turns out there are tests doing two write in a row without any
read in the middle. It was throwing an exception when this
happened.
This patch makes the server to maintain a queue of messages instead. It
doesn't throw errors anymore.
Bug: chromium:1109194
Change-Id: I25f8dc2c2a032781a4527a0994e94e46f85e12c9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2317865
Reviewed-by: Pâris Meuleman <pmeuleman@chromium.org>
Commit-Queue: Arthur Sonzogni <arthursonzogni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#791901}
--
wpt-commits: f4927f9d8023557d15cddd8d9eb40ac079bf697e
wpt-pr: 24740
Automatic update from web-platform-tests
[css-lists] Fix marker not correctly updated in legacy layout
If a list item contains a block box, the layout tree is initially like:
LayoutListItem {LI}
LayoutBlockFlow (anonymous)
LayoutListMarker {::marker}
LayoutBlockFlow {P}
LayoutText #text
However, an outside ::marker should appear in the same line as the first
non-marker text of the list item. Therefore, it's changed into:
LayoutListItem {LI}
LayoutBlockFlow (anonymous)
LayoutBlockFlow {P}
LayoutListMarker {::marker}
LayoutText {#text}
But this was not happening when dynamically toggling list-style-type
between 'none' and some other value. The reason was that there was a
'NormalChildNeedsLayout()' condition, which would hold the 1st time, but
in following layouts it would be false.
This patch just removes that condition.
Bug: 1107783
TEST=external/wpt/css/css-lists/change-list-style-type-002.html
Change-Id: Ic82f1a474b0a0ddf2f2aab8a60afac299aa6e2d1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2310359
Reviewed-by: Koji Ishii <kojii@chromium.org>
Commit-Queue: Oriol Brufau <obrufau@igalia.com>
Cr-Commit-Position: refs/heads/master@{#791857}
--
wpt-commits: 79331ece5c30ca14a3bcc1674d4efe19f0657754
wpt-pr: 24692
Automatic update from web-platform-tests
Remove use of setTimeout in ref test for web-animations
Use requestAnimationFrames instead of setTimeouts in the ref tests to
accommodate variability in timing. The background color is animated
between green and red to better align with reftest guidelines.
The test to slow down the animation was removed since it would be
essentially equivalent to the zero playback rate test after the
refactoring.
Bug: 1108342, 1064127
Change-Id: Ie933b68cf45caabfa70fbe5deb881e120ea5e28a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2313157
Reviewed-by: Xida Chen <xidachen@chromium.org>
Commit-Queue: Kevin Ellis <kevers@chromium.org>
Cr-Commit-Position: refs/heads/master@{#791833}
--
wpt-commits: 5042344eb8042c6a5be8815add424e45a880f0fc
wpt-pr: 24727
Automatic update from web-platform-tests
[COOP report] Add support for report-only
This adds the support for report-only COOP triggered Browsing context
group switches.
During navigation, besides computing whether COOP triggers a BCG switch,
this also computes if any of the report-only policies would also trigger
a switch if they were effective, the resulting values are stored in
booleans within |CrossOriginOpenerPolicyStatus|.
This booleans are then used to trigger the navigation reports, and in a
follow up to trigger access reports.
Explainer [WIP]:
https://github.com/camillelamy/explainers/blob/master/coop_reporting.md
Specification [WIP]:
https://github.com/whatwg/html/pull/5518
Bug: 1099208
Change-Id: I2cb66ec5cdcd9d5b4658c28f0608bc6b52d0da6b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2241529
Commit-Queue: Pâris Meuleman <pmeuleman@chromium.org>
Reviewed-by: Camille Lamy <clamy@chromium.org>
Reviewed-by: Arthur Hemery <ahemery@chromium.org>
Auto-Submit: Pâris Meuleman <pmeuleman@chromium.org>
Cr-Commit-Position: refs/heads/master@{#790781}
--
wpt-commits: 969178e38dd571ff68b74322430ecdf572abe2a3
wpt-pr: 24552
Automatic update from web-platform-tests
[css-pseudo] Support text-transform in ::marker
In https://github.com/w3c/csswg-drafts/issues/4206, the CSSWG resolved
that the 'text-transform' property should apply to ::marker, and that
by default it should be set to 'none' in the UA stylesheet.
This patch implements these changes.
Note that setting it to 'none' is an observable change in LayoutNG,
since it prevents the ::marker from inheriting the 'text-transform' of
the list item. This was considered a better behavior by the CSSWG, and
it's the behavior that legacy layout already had.
LayoutListMarker, used for ::marker with 'content: normal' in legacy
layout, doesn't implement the marker text as actual text. Therefore,
'text-transform' has no effect.
Bug: 1031667
TEST=external/wpt/css/css-pseudo/marker-default-styles.html
TEST=external/wpt/css/css-pseudo/marker-text-transform-default.html
TEST=external/wpt/css/css-pseudo/marker-text-transform-uppercase.html
TEST=external/wpt/css/css-pseudo/parsing/marker-supported-properties-in-animation.html
TEST=external/wpt/css/css-pseudo/parsing/marker-supported-properties.html
TEST=http/tests/devtools/elements/styles-2/pseudo-elements.js
Change-Id: I2c2b554f521139fabb502856fedc6230aa45c787
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2315872
Reviewed-by: Koji Ishii <kojii@chromium.org>
Commit-Queue: Oriol Brufau <obrufau@igalia.com>
Cr-Commit-Position: refs/heads/master@{#791815}
--
wpt-commits: 22901727d52297378d44b217af0b4c06d5b0a484
wpt-pr: 24725
Automatic update from web-platform-tests
Add wpt test for font feature precedence for 'dlig'.
The new Lato-Medium-Liga font is based on Lato-Medium 3.002 available
at https://github.com/latofonts/lato-source/tree/master/fonts/static/TTF
It is subset with hb-subset with "fiststfi " and the 'fi' ligatures
moved back to 'liga' (version 3.002 moved them to 'dlig'). The crossbar
of 't' in the 'st' ligature was also shorted to ensure that it would not
cover a normal 't'.
The Lato-Medium currently available to the tests is version 2.015 and
lacks a mapping from U+fb06 to the 'st' ligature glyph which was only
accessible through the 'dlig' feature. Lato-Medium 3.002 added this
mapping (required for this new test) but also moved the 'fi' ligatures
to 'dlig' and some existing tests depend on the 'fi' ligatures being in
'liga'.
This tests https://github.com/w3c/csswg-drafts/pull/3430
Change-Id: Iaf0ff1f99f3ec282bacb0e3cb44ede27d47fac5a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1294874
Reviewed-by: Dominik Röttsches <drott@chromium.org>
Commit-Queue: Ben Wagner <bungeman@chromium.org>
Cr-Commit-Position: refs/heads/master@{#791807}
--
wpt-commits: f76fb70a30ce97c33425b30e4654ab4b5bce4fd1
wpt-pr: 14470
Ian Kilpatrick
Marijn Kruisselbrink
David Grogan
Ziran Sun
Philipp Hancke
Stephen McGruer
Gérard Talbot
moz-wptsync-bot
Oriol Brufau
Xianzhu Wang
Robert Ma
Guido Urdaneta
arthursonzogni
Ms2ger
autofoolip
Tim Schumann
Chris Harrelson
Olga Gerchikov
Dylan Cutler
George Steel
Adithya Srinivasan
Kevin Ellis
Pâris MEULEMAN
Ben Wagner