The preference "test.events.async.enabled" controls the usage of
widget events in general, but Marionette doesn't support all of
them yet. To get widget events enabled for web-platform tests
those unsupported event types need to still synthesize the event.
Differential Revision: https://phabricator.services.mozilla.com/D188296
The filterBySource was looping over the whole list once again
but worse, was duplicating the array once more.
Same for filterByUniqLocation.
Finally, once we removed these two last filtering methods... we could prevent the creation of the whole "positions"
intermediate object and craft directly what we aim to store in redux.
Differential Revision: https://phabricator.services.mozilla.com/D189951
We can avoid the creation of an handful of object creations by passing
the "breakpoint positions" objects directly to the Source Map Loader.
(avoid convertToList)
We can also tune `SourceMapLoader.getOriginalLocations` output to better
fit the requirements of the main thread.
Differential Revision: https://phabricator.services.mozilla.com/D189950
This code is a bit hard to follow, so let's add some comments.
Also `columns = [...new Set([...existing, ...columns])];` is probably very bad for performance.
It clones existing and columns into a third array, whose content is cloned into a Set,
which is ultimately cloned into the final array.
Differential Revision: https://phabricator.services.mozilla.com/D189948
This wasn't a problem before bug 1843044 because for dark mode we
ignored HCM (as it was a lightweight theme), and because on Windows
HCM implies light mode.
Differential Revision: https://phabricator.services.mozilla.com/D190185
Many of these files are now in unmodified form. These changes produce
no changes when running:
./mach python python/mozbuild/mozbuild/gn_processor.py dom/media/webrtc/third_party_build/gn-configs/webrtc.json
Depends on D190104
Differential Revision: https://phabricator.services.mozilla.com/D190105
I'm hesitant to use RenderBlocking in the sense that... well, we know we
are not going to get to fetch it between rendering. Maybe this?
Differential Revision: https://phabricator.services.mozilla.com/D190107
We hit some limitations of the globbing system in exclude file, which
relies on glob.iglob and skips dotfiles...
As a side effect update the license to the one actually shipped with
libfuzzer.
Differential Revision: https://phabricator.services.mozilla.com/D190068
Automatic update from web-platform-tests
Make navigation-api/navigate-event/ resilient to starting with extra NavigationHistoryEntries
When WPT are run in-browser or via wptrunner, there may be additional
session history entries when the test starts, so the tests shouldn't
assume that they start with navigation.entries().length == 1 or
navigation.currentEntry.index == 0.
This is a workaround for https://github.com/web-platform-tests/wpt/issues/33590.
Change-Id: I978d492d5d31d110716d6a1d84732aff792aca14
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4904223
Reviewed-by: Domenic Denicola <domenic@chromium.org>
Auto-Submit: Nate Chapin <japhet@chromium.org>
Commit-Queue: Nate Chapin <japhet@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204863}
--
wpt-commits: 4cd570dc80fec7b17cad121e7ac866badf62df38
wpt-pr: 42233
Automatic update from web-platform-tests
gamepad: Check that the DocumentLoader is not null
Gamepad timestamps are reported relative to the navigationStart
timestamp, accessed from the DocumentLoader's DocumentLoadTiming.
In some scenarios the document may have been detached before the
NavigatorGamepad is initialized, causing the DocumentLoader to be
nullptr.
This CL checks that the DocumentLoader is non-null, and uses the
current time in place of navigationStart if it is null.
Bug: 1483777
Change-Id: Ib3efcf3a612cd4c1c543218fa5fb8568c7129dcf
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4899018
Reviewed-by: Alvin Ji <alvinji@chromium.org>
Commit-Queue: Matt Reynolds <mattreynolds@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204813}
--
wpt-commits: 3ae2cb8c5b5ba61809da15318821a6c1fe54829f
wpt-pr: 42227
Automatic update from web-platform-tests
Optimize several things about DOM Parts [1/2]
See [1] for details. This CL makes these changes:
- Stop maintaining “live” parts lists across DOM mutations. This is
controlled by a new feature flag, DOMPartsAPIActivePartTracking,
that when disabled, short-circuits all of the part list maintenance
code during DOM mutations. This means that if DOM nodes that are
connected to parts are moved around in the DOM, including being
removed from the document, the cached parts list does not change.
- Update ToV8HelperSequence to work directly with Deques, and get
rid of getParts() code to convert to a HeapVector.
- Optimize ToV8HelperSequence to use iterators instead of array
indexing, which should also speed up all sequence bindings. (TBD)
- Change from FrozenArray<Part> to sequence<Part>. This eliminated
the overhead of Object::SetIntegrityLevel(), which was ~29% of the
total getParts() time.
- Add a fast path for PartRoot.replaceChildren when all children
are being removed.
- Inline IsValid().
- Call InsertBefore() instead of before() to avoid some extra
validity checks. This required making ConvertNodeUnionsIntoNode
public.
- Add PartRoot.getPartNode(n), which allows direct access to a
single Part's Node, to avoid wrapping up the entire parts list.
[1] https://docs.google.com/document/d/1wSbBV6tS-KLlHWxaK_FmEJFOwkY-pb8_YDoFJgDrbMg/edit
Bug: 1453291
Change-Id: I3df15b2a552fe9c2793a7f7f2095d632bcea8326
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4878048
Reviewed-by: Joey Arhar <jarhar@chromium.org>
Commit-Queue: Mason Freed <masonf@chromium.org>
Auto-Submit: Mason Freed <masonf@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204772}
--
wpt-commits: 27ccadd96797aec13652caaeaab3a428c06ab2d3
wpt-pr: 42083
Automatic update from web-platform-tests
Parse and use different reporting endpoints per feature in permissions policy
Change-Id: I1cc8a1ea185500d44ddb2718803c22f2038f66c6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4473627
Reviewed-by: Kouhei Ueno <kouhei@chromium.org>
Commit-Queue: Ian Clelland <iclelland@chromium.org>
Reviewed-by: Ken Buchanan <kenrb@chromium.org>
Reviewed-by: Ari Chivukula <arichiv@chromium.org>
Reviewed-by: Adam Rice <ricea@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204761}
--
wpt-commits: 592f7b1cf7c9fb40c514940950ae2120339d4251
wpt-pr: 42252
Automatic update from web-platform-tests
[FLEDGE]: Subdivide WPT tests.
FLEDGE tests tend to be fairly slow, since they can spawn a bunch of
processes. This CL subdivides most FLEDGE WPT tests into sets of 5
(one larger file into sets of 10), which should hopefully reduce flake
on the bots.
Fixed: 1481290
Change-Id: If708b33d0713e783cc53f014d9fbfe127ec1dc9d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4909551
Reviewed-by: Maks Orlovich <morlovich@chromium.org>
Commit-Queue: Matt Menke <mmenke@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204664}
--
wpt-commits: f6d8468144f5f9f8fe2e6b8ecdb1846ce08c3700
wpt-pr: 42312
Automatic update from web-platform-tests
Add test that "maintain-framerate" does what it says.
Adds a test that measures the bandwidth, cuts the bandwidth in half,
and then measures that framerate is maintained, but resolution is not.
Also adds more entropy to the "getNoiseStream()" video signal; the old
version generated a video stream of 8 Kbits/second in 640x480.
Bug: none
Change-Id: Iefe0a44b9686f1bb4bdd13ab63922c3d460c775e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3117645
Reviewed-by: Philipp Hancke <phancke@microsoft.com>
Commit-Queue: Harald Alvestrand <hta@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204584}
--
wpt-commits: 1af7ffeb926a3644c33bcbf7ab0a1bb6c2f0a824
wpt-pr: 30173
Automatic update from web-platform-tests
Fix WPT CSS Counters test
As per:
https://drafts.csswg.org/css-contain/#example-6932a400
and
https://github.com/w3c/csswg-drafts/issues/9116
counter() and counters() can cross the style containment boundary,
so the result should be 13 from the body's counter-reset.
Bug: 990657
Change-Id: Iee313a5ed6f487865b1b7c754aa45b736f6b8f15
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4836162
Reviewed-by: Rune Lillesveen <futhark@chromium.org>
Commit-Queue: Daniil Sakhapov <sakhapov@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204571}
--
wpt-commits: dcb79ca1856efaa5ac6b455d1f96f63e3038e663
wpt-pr: 41799
Automatic update from web-platform-tests
Update dependencies in HTML idlharness test
--
wpt-commits: 039581da9bf12f29ecd4a9915a04601def843278
wpt-pr: 42284
Automatic update from web-platform-tests
Ignore complex relative expression in css atan2()
Currently, we use the double values if the parameter is relative type.
However, if the parameter is an expression that includes the relative
value it cannot be resolved as a simple double value.
It causes a crash because trying to use the value that is not
resolved.
Therefore, parsing a failure if ResolvedUnitType() function returns
kUnknown which means complex relative expression at the parsing time
until supporting "Involving relative values in Non-Length CSS Math
Expressions" [1]
[1] https://crbug.com/1436612
Bug: 1476829
Change-Id: If920388de96d2838eccf4d76fecb050e414d29dd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4835559
Reviewed-by: Anders Hartvoll Ruud <andruud@chromium.org>
Commit-Queue: Seokho Song <seokho@chromium.org>
Reviewed-by: Xiaocheng Hu <xiaochengh@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204500}
--
wpt-commits: 78701298a872dc884f7bd341108d0a5ff8c41bf2
wpt-pr: 42050
Automatic update from web-platform-tests
Update webcodecs decoder test as error callback is now executed before flush promise callbacks (#42285)
--
wpt-commits: 68cfc233182799e3c4edc2e6d7c43b7f795c1a2f
wpt-pr: 42285
Automatic update from web-platform-tests
Paint background invalidation when deleting the last element of TableRow
When the last element in a TableRow is deleted, it does not trigger
a redraw of the background,causing the background to persist.
After removing the last element of a row in a Table,
the size of the TableRow remains unchanged because it does not affect
the layout of the preceding elements.
Therefore, caching is used during rendering.
We should proactively trigger a redraw of this row when deleting
the last element of a table row.
Bug: 1487554
Change-Id: Ia45f5abf8ccd54781e8f0c1de04aebfa5cad552c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/4900123
Reviewed-by: Rune Lillesveen <futhark@chromium.org>
Reviewed-by: Xianzhu Wang <wangxianzhu@chromium.org>
Commit-Queue: Xianzhu Wang <wangxianzhu@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1204290}
--
wpt-commits: 6447f83eae831c2470b4e8eb5cd6af6fbbe0103e
wpt-pr: 42217
