Automatic update from web-platform-tests
Implement LCP type DataURI for videos.
Videos of data URI src is supported in HTML5. It is also parsed by GURL
parser and the information is passed down to the ImageResourceContent
class.
This CL just wires it up.
Bug: 1383959
Change-Id: I2fb2edad66c5f0d3bfa700c7cf67207ba898fc23
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4033006
Commit-Queue: Hao Liu <haoliuk@chromium.org>
Reviewed-by: Yoav Weiss <yoavweiss@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1080327}
--
wpt-commits: 7005b1f855ffa3443d30ca3ae2d2084203f68b75
wpt-pr: 37087
Automatic update from web-platform-tests
Ensure Soft Navigation LCP, FCP and FP fire more than once
There were a couple of issues preventing paint entries from firing after
the first soft navigation is detected:
1) The related flags were reset only once.
2) Image paint flags were not reset at all.
This CL fixes those two issues.
There's another issue [1] this work revealed, which would wait for a
followup.
[1] https://crbug.com/1397720
Bug: 1395507
Change-Id: Ib5aa9f240c7d67f934315a2643ccc022ad4ced72
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4083642
Reviewed-by: Ian Clelland <iclelland@chromium.org>
Commit-Queue: Yoav Weiss <yoavweiss@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1080310}
--
wpt-commits: c410fadaa5fe1d9401d10287c21143410f5ea6ca
wpt-pr: 37369
Automatic update from web-platform-tests
webrtc wpt: assert that setParameters implements the transaction model
as described in
https://w3c.github.io/webrtc-pc/#dom-rtcrtpsender-setparameters
BUG=None
Change-Id: Ie30f6fc3bafadff66cc9f4298984d57946440975
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4083428
Reviewed-by: Harald Alvestrand <hta@chromium.org>
Commit-Queue: Philipp Hancke <phancke@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#1080285}
--
wpt-commits: 749e04ef7be338b84a778b200a814998fd78948a
wpt-pr: 37367
GCRuntime::markAllDelayedChildren can be called during gray marking if we OOM
there while expanding the mark stack. It doesn't matter since the mark stack is
empty and processDelayedMarkingList will save and restore the original color.
I missed this second testcase in the bug. It's a different issue than the one
in the description and should really have been given its own bug.
Differential Revision: https://phabricator.services.mozilla.com/D164374
This assertion is going off because it thinks we souldn't be accessing a cell's
zone off-thread. In fact we are marking and the mutator is not running so this
is fine. The problem is we didn't set GC use to marking for the background mark
task (not parallel marking, this does single threaded marking off-thread during
sweeping).
Differential Revision: https://phabricator.services.mozilla.com/D164320
This patch accomplishes a couple different things in one go.
about:welcome has a feature to allow switching screens by history
navigation. But the screen transitions don't play in that case. There
was also an issue (see bug 1804098) where rounding inconsistencies could
cause 1px gaps to appear in the animated progress bar, because of the
use of transform. By converting the progress bar into a single element
and just modulating its width, we can stop using transform and also
enable transitions for history navigation. Technically, this also allows
us to advance by more than 1 screen without messing up the animations,
though this is not currently used. Incidentally, it simplifies the CSS.
Differential Revision: https://phabricator.services.mozilla.com/D164310
Use asynchronous methods to retrieve or generate page thumbnails. Also
add explicit handling for unloaded browsers without stored page
thumbnails. In that case, give the preview frame a fallback style
instead of showing a broken img or an empty black canvas.
Differential Revision: https://phabricator.services.mozilla.com/D159672
clang is built using the sysroot, and not using the sysroot to build the
afl-llvm-pass plugin can lead to ABI differences that lead to crashes.
Differential Revision: https://phabricator.services.mozilla.com/D163905
Automatic update from web-platform-tests
Screen Capture: Fix getdisplaymedia-framerate.https.html test
This CL fixes a Screen Capture test that uses getVideoPlaybackQuality by
making sure that mozVideoFrames (not supported by Chromium) is only used
if supported.
Bug: 1384645
Change-Id: I0e163b48370e7f25e84bf8797b656b5f1aef5865
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4084761
Commit-Queue: Fr <beaufort.francois@gmail.com>
Reviewed-by: Tove Petersson <tovep@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1080274}
--
wpt-commits: 4047d18232b73b05a6f15667316398c66ed421b5
wpt-pr: 37371
Automatic update from web-platform-tests
Screen Capture: Fix getdisplaymedia-framerate.https.html test
This CL fixes a Screen Capture test that uses getVideoPlaybackQuality by
making sure that mozVideoFrames (not supported by Chromium) is not used if getVideoPlaybackQuality.totalVideoFrames is equal to 0.
Bug: 1384645
Change-Id: I194bdbbb03701a035f5e4b1f0591e97bd3149cfb
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4080391
Reviewed-by: Tove Petersson <tovep@chromium.org>
Commit-Queue: Fr <beaufort.francois@gmail.com>
Cr-Commit-Position: refs/heads/main@{#1080221}
--
wpt-commits: 21b6094665a8eb23769dac57f980848dc836ec2a
wpt-pr: 37368
Automatic update from web-platform-tests
[mathml] Add tests for MathML and floated elements
This adds tests to verify how the float property interacts with
MathML layout. It also adds some logic to ignore the float
property inside math display.
Bug: 6606
Change-Id: I415d049b9783d699f4f2ceb3ca15d4c67243950d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4055563
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Commit-Queue: Frédéric Wang <fwang@igalia.com>
Cr-Commit-Position: refs/heads/main@{#1080219}
--
wpt-commits: b3d201bfc4c0f8c9b42b8fefba257b4d9324af08
wpt-pr: 37162
Automatic update from web-platform-tests
LayoutNG: Fix PositionedObjects for fieldset content box
This is a follow-up of crrev.com/c/3767050.
It made ComputeIsAbsoluteContainer(old_style) not workable because the
function refers to the current style of the parent fieldset. We failed
to clear PositionedObjects() for fieldset content boxes because of it.
LayoutObject has a cached value for ComputeIsAbsoluteContainer(), and
we don't need to call ComputeIsAbsoluteContainer() for old_style now.
Bug: 1395688
Change-Id: I4d3da0c94d568a9935da01ae5fd0b86429c0ae7f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4082352
Auto-Submit: Kent Tamura <tkent@chromium.org>
Reviewed-by: Koji Ishii <kojii@chromium.org>
Commit-Queue: Koji Ishii <kojii@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1080181}
--
wpt-commits: d36c4bceb008ac60a8ee9d0d204f4a2bbbab6377
wpt-pr: 37365
Automatic update from web-platform-tests
Fix WPT for selection changes on canceled mousemove.
A canceled mousemove should not suppress the following selectionchange
event. The selectionchange event after mousedown is not relevant
here.
Bug: 346473
Change-Id: I48f776b68cda4b2294fdd973bf45c2e88d1d47ff
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4081610
Reviewed-by: Kevin Ellis <kevers@chromium.org>
Commit-Queue: Mustaq Ahmed <mustaq@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1079979}
--
wpt-commits: ddb55e65764b0be1154b45fa953001d2acace941
wpt-pr: 37358
Automatic update from web-platform-tests
Trigger layout explicitely in cleanup for contain-intrinsic-size/auto-006.html (#37335)
In WebKit, it does not always trigger layout when JS changes CSS value.
In auto-006.html, the cases need to trigger layout in cleanup(), otherwise the previous last remembered size might affect the next case.
So add checkSize to trigger layout explicitly in cleanup().
--
wpt-commits: 3f118ecfbe926d782cbfb726c609f80cb57ae897
wpt-pr: 37335
Automatic update from web-platform-tests
Fix resolving of SMIL intervals in the past
If the only active interval of a timed element was in the past when the
first interval was resolved, no interval would be resolved. This in turn
could mean that an animation should still apply after the active
duration ended (i.e because of fill=freeze) did not.
When computing the first interval for a timed element, then store/update
interval state after each resolved interval that is reject (because it's
in the past). This means that the initial interval state will reflect
the same state that it would get from essentially fast-fowarding through
the intervals. If an unresolved interval ends up being returned from
ResolveInterval() - like in the bug in question - then it will move the
previously resolved interval (`interval_`) into `previous_interval_`.
Thus there's no need to update `previous_interval_` in the resolve loop.
Setting `is_waiting_for_first_interval_` means that the correct active
state will be computed by `DetermineActiveState()`.
Fixed: 1395605
Change-Id: Iac185713f298608bd1d3337425fe32188a7670fe
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4079445
Commit-Queue: Philip Rogers <pdr@chromium.org>
Reviewed-by: Philip Rogers <pdr@chromium.org>
Auto-Submit: Fredrik Söderquist <fs@opera.com>
Cr-Commit-Position: refs/heads/main@{#1079792}
--
wpt-commits: 9d2a7aff2f25e0c18f4906d945284e0beef4bf23
wpt-pr: 37329
Automatic update from web-platform-tests
Fix timeout in WebKit for lock-sandboxed-iframe.html (#37301)
Fix timeout in WebKit for lock-sandboxed-iframe.html by adding a try / catch
around the exitFullscreen() call, in case the promise gets rejected.
--
wpt-commits: 171909d69731612d3023b830e3fd7406439d5fd1
wpt-pr: 37301
Automatic update from web-platform-tests
Allow the empty string as a referrer policy in speculation rules
We now consider the empty string valid and treat it as if the
referrer_policy key had been omitted, instead of discarding the rule.
Bug: 1355146
Change-Id: Icded81e0b71a59d07c3369da8e8bf6aeeb2578d5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4081231
Auto-Submit: Kevin McNee <mcnee@chromium.org>
Commit-Queue: Adithya Srinivasan <adithyas@chromium.org>
Reviewed-by: Adithya Srinivasan <adithyas@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1079471}
--
wpt-commits: 65d8e9cff31b3acebd113d90cadefc228fc00efc
wpt-pr: 37333
Automatic update from web-platform-tests
Relax same-site prefetch referrer policy check
Applying the sufficiently-strict referrer policy restriction to all
prefetches was found to break legitimate same-site use cases. We now
only enforce this for cross-site prefetches. This also makes the logic
consistent between prefetch and prerender.
Bug: 1379846, 1355146
Change-Id: I945cc4f584af83e42cf80f146ee96823f2ada27a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4077268
Auto-Submit: Kevin McNee <mcnee@chromium.org>
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Commit-Queue: Jeremy Roman <jbroman@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1079453}
--
wpt-commits: 7d839337f49d4c43ff5e00ad93b38145e06572b9
wpt-pr: 37326
Automatic update from web-platform-tests
[anchor-position] Implement scroll-aware position fallback
This patch brings scroll offset into position fallback algorithm to
decide which fallback position to use, and also calculates info to
help decide when to invalidate the current fallback position.
Bug: 1371217
Change-Id: I5256b1422e6f48cfbb1c70e6f47c7e65362d3496
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3995704
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Commit-Queue: Xiaocheng Hu <xiaochengh@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1079447}
--
wpt-commits: 1b1324fb49e6ff343ae371659f4d2d6b4d1474e6
wpt-pr: 36756
