A function marked `isAsmJSNative` is either the outer module function or an actual
exported asm.js function. The module function can't be disassembled.
Differential Revision: https://phabricator.services.mozilla.com/D225357
Collecting the full history updates from the parent process until the partial history colletion is introduced for SHIP on Android.
Differential Revision: https://phabricator.services.mozilla.com/D221710
IndexedDB still calls QuotaManager::EnsurePersistentOriginIsInitializedInternal
from its idle database maintenance. However, we want to get rid of that so the
method can be made private. The idle database maintenance should only rely on
QuotaManager::OpenStorageDirectory method to do all necessary initialization.
Before we can do that, we need a method dedicated for initialization of entire
persistent storage.
Differential Revision: https://phabricator.services.mozilla.com/D195653
Task inner structure contains an #ifdef from a variable defined in
nsThreadUtils.h.
Otherwise we may end up with different compilation unit have different
vtable (odr violation).
Differential Revision: https://phabricator.services.mozilla.com/D224754
Automatic update from web-platform-tests
[fenced-frames] Migrate WebAuthn test
Migrate a fenced frames WebAuthn test to run on headless chrome with
chromedriver instead of the legacy content shell runner. The legacy
content shell runner will stop supporting the WebAuthn testdriver API,
soon.
Fix the test to set up the virtual authenticator environment on the top
frame instead, since testdriver should be invoked from the same window
hosting the testharness.
Bug: 372169469
Change-Id: I6e1b296fe86ee4087da2a5a63868e18a6b440984
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5920770
Auto-Submit: Nina Satragno <nsatragno@chromium.org>
Reviewed-by: Liam Brady <lbrady@google.com>
Commit-Queue: Liam Brady <lbrady@google.com>
Cr-Commit-Position: refs/heads/main@{#1367535}
--
wpt-commits: 5611c150825562522da2d6be67b929e399211c35
wpt-pr: 48579
Automatic update from web-platform-tests
[wptrunner] Cycle testdriver event loop when testharness isn't loaded (#48582)
Workaround for https://crbug.com/340662810 that seems to work
without an increase in test runtime or regressions
(https://crrev.com/c/5698330/9).
--
wpt-commits: 280d04eaabed08d430d12c8af971b0b56f32310f
wpt-pr: 48582
Automatic update from web-platform-tests
Rename & refactor invoker commands to reflect new spec decisions.
Invoker commands now live solely on HTMLButtonElement, and have been
dropped from HTMLInputElement. This was discussed and resolved + minuted
in Jul 18 WHATNOT meeting:
https://github.com/whatwg/html/issues/10471#issuecomment-2237135334)
Words must be separated with dashes. This was discussed and resolved +
minuted in Jul 18 WHATNOT meeting:
https://github.com/whatwg/html/issues/10471#issuecomment-2237135334)
"Custom" commands now require a double dash (`--`) prefix. This was
discussed & resolved + minuted in the Jul 25 OpenUI meeting:
https://github.com/openui/open-ui/issues/969#issuecomment-2251153158
In addition, CommandEvent#invoker is now renamed to CommandEvent#source
as discussed and resolved + minuted in Jul 18 WHATNOT meeting:
https://github.com/whatwg/html/issues/10471#issuecomment-2237135334)
This CL updates the tests & implementation to follow these changes. The
changes can also be found in the spec PR which reflects this updated
status:
https://github.com/whatwg/html/pull/9841
Bug: 40284894
Change-Id: Iefa91a1736867239a21f8717aee1fcaaa54fdfd0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5839601
Reviewed-by: Mason Freed <masonf@chromium.org>
Reviewed-by: Luke <lwarlow@igalia.com>
Commit-Queue: Keith Cirkel <chromium@keithcirkel.co.uk>
Cr-Commit-Position: refs/heads/main@{#1367705}
--
wpt-commits: 5df7a2c02bb50efa0368910fd49242c08def7a5d
wpt-pr: 48581
Automatic update from web-platform-tests
Add implicit anchor relationship for popover invokers
When PopoverAnchorRelationships is enabled, include popover invokers
as "implicit anchor elements" for anchor positioning.
This also adds a test of the behavior when nesting <select> within
other top layer elements, and vice versa.
Bug: 364669918
Change-Id: Ie1d362a41e3aec72d62dbfe378e65ea9a262c324
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5846489
Auto-Submit: Mason Freed <masonf@chromium.org>
Commit-Queue: Joey Arhar <jarhar@chromium.org>
Reviewed-by: Joey Arhar <jarhar@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1367631}
--
wpt-commits: 39246a878f6b2e8cd4387896d05318f9e0e1e85a
wpt-pr: 48580
Automatic update from web-platform-tests
Shadow trees: more involved dispatch tests
In particular, this adds tests for when the target get moved out of the shadow tree or into a shadow tree by an event listener.
Co-authored-by: Chris Dumez <cdumez@apple.com>
--
wpt-commits: b330bd718452c3dc9b7f6826ba62d51bf81d5f4d
wpt-pr: 48577
Automatic update from web-platform-tests
Fix check failure at LineBreaker::HandleTextForFastMinContent
The DCHECK_EQ fails, because sometimes the decoration width of elements
with `box-decoration-break` is zero, so `position_` will be
equal to `indent`, and the value of `should_break_at_first_opportunity`
will be false. `has_cloned_box_decorations_` is true. So the DCHECK_EQ
fails.
Bug: 362530623
Change-Id: I4523c16089326340c4bab88ae3b674ccb0bdd9ee
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5873209
Reviewed-by: Morten Stenshorne <mstensho@chromium.org>
Reviewed-by: Koji Ishii <kojii@chromium.org>
Commit-Queue: Morten Stenshorne <mstensho@chromium.org>
Auto-Submit: cathie chen <cchen@igalia.com>
Cr-Commit-Position: refs/heads/main@{#1367328}
--
wpt-commits: f28c47aeda5fe8c5722e176ba97a3980e507c0c1
wpt-pr: 48565
Automatic update from web-platform-tests
webnn: utilize constant nodes in DirectML backend
This CL utilizes `DML_GRAPH_NODE_TYPE_CONSTANT` introduced in
`DML_FEATURE_LEVEL_6_2` for constant scalar inputs, which allows CPU
memory to be directly consumed during graph compilation.
Change-Id: I468e600e0b1dae59a0ab017ef7fe86b5e0e00c59
Bug: 366284850
Cq-Include-Trybots: luci.chromium.try:win11-blink-rel, mac14.arm64-blink-rel, mac14-blink-rel, mac15.arm64-blink-rel, mac15-blink-rel, linux-blink-rel
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5859935
Reviewed-by: Rafael Cintron <rafael.cintron@microsoft.com>
Commit-Queue: Shiyi Zou <shiyi.zou@intel.com>
Reviewed-by: ningxin hu <ningxin.hu@intel.com>
Auto-Submit: Shiyi Zou <shiyi.zou@intel.com>
Cr-Commit-Position: refs/heads/main@{#1367255}
--
wpt-commits: 8903062b181d2fdf406e7b74083f13af59012c15
wpt-pr: 48568
Automatic update from web-platform-tests
[line-clamp] Hide from paint non-fixed-height abspos after clamp
With the `continue: collapse` variant of `line-clamp`, absolutely
positioned boxes that are after the clamp point in the layout tree
should be hidden from paint. However, abspos with an indefinite block
size were not always being hidden.
The reason for this was that for such abspos,
`OutOfFlowLayoutPart::CalculateOffset` performs an initial layout,
which can end up becoming the final layout of the abspos, which does
not happen for the fixed-height abspos that we had tested previously.
As it turns out, that layout was not checking the `NodeInfo`'s
`is_hidden_for_paint` field to determine whether to hide the fragment.
This change fixes that.
Bug: 40336192
Change-Id: I5876c762d0d9957ea5785660c73528b83de99eb9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5923076
Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org>
Commit-Queue: Andreu Botella <abotella@igalia.com>
Cr-Commit-Position: refs/heads/main@{#1367195}
--
wpt-commits: 0ca2856475e2d067f7f2523ffb30de965c28d14a
wpt-pr: 48564
Automatic update from web-platform-tests
[reporting][wpt] Isolate stashed reports
Several `reporting/` tests use the same wptserve stash entry (report ID
`d0d517bf-891b-457a-b970-8b2b2c81a0bf` [0]). If at least two such tests
are unlucky enough to run in parallel, the first one to poll `report.py`
will cannibalize all of the reports and get too many, while the others
will get too few.
Fix this by regenerating different IDs for all but one of the tests. For
the external WPTs that embed on `middle-frame.https.sub.html`, plumb the
report ID to `same-origin-report.https.sub.html` as a query parameter,
where it can be substituted into the `Reporting-Endpoints` response
header [1].
Continue using hardcoded IDs instead of randomly generated ones so that
reports can be matched with their tests.
[0]: https://source.chromium.org/search?q=d0d517bf-891b-457a-b970-8b2b2c81a0bf%20f:wpt%20&ss=chromium%2Fchromium%2Fsrc
[1]: https://web-platform-tests.org/writing-tests/server-pipes.html#sub
Bug: 338470458
Test: run_web_tests.py external/wpt/reporting wpt_internal/reporting --fully-parallel -j 36
Change-Id: I46d4d9de954d853615b067725a136a65a6df622b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5922396
Reviewed-by: Ian Clelland <iclelland@chromium.org>
Commit-Queue: Jonathan Lee <jonathanjlee@google.com>
Cr-Commit-Position: refs/heads/main@{#1367116}
--
wpt-commits: 3e0163f8505cac8a369aab811179cc9dc0f60ebe
wpt-pr: 48570
Use some kind of prefix table to avoid storing redundant string. Also
Avoid storing an array of enum when there's a direct mapping between
enum values and loop index.
This result in a reduction of 352 byte in the resulting binary, so this
may not be worth the effort, but it was great fun to write that patch.
Differential Revision: https://phabricator.services.mozilla.com/D224494
Depends on D224505
Cross browser tests are added for basic image cache scenarios.
An additional mozilla specific wdspec test is added to cover scenarios where the behavior is implementation defined.
Differential Revision: https://phabricator.services.mozilla.com/D224506
Previously when the keyboard would be showing the entire screen will first resize to the end height
after which the keyboard will be animated over an empty content.
With these changes we postpone the resize to after the keyboard animation is finished to ensure a
smooth animation over the web content instead of over empty space.
Differential Revision: https://phabricator.services.mozilla.com/D223586
Previously when the keyboard would be showing / hiding the entire screen will first resize to
the end height after which the keyboard will be animated.
This meant just one reflow for the engine view but a choppy layout animation.
With the new changes we'll ensure the bottom toolbar is animated in sync with the keyboard
and the engine view is resized only once
- before hiding the keyboard to ensure that it will then nicely reveal the web content
- after showing the full keyboard to ensure it gradually covered web content with no UI issue
Differential Revision: https://phabricator.services.mozilla.com/D223585
Previously, the event listeners of a previous test run would stay,
contributing to the set of assertions of the current test run.
Differential Revision: https://phabricator.services.mozilla.com/D223595
The important change (and only change needed to fix the bug) is to remove `EngineViewClippingBehavior`
if already set which would update how the engine view is displayed to account for the dynamic toolbars
if in PWAs we don't show toolbars.
Since we want the same functionality for configuring toolbars and the engine view when toolbars
should be hidden in a PWA or when entering fullscreen and also when toolbars should be shown again
we'll use now same methods in both places which encapsulate this potentially brittle code.
Differential Revision: https://phabricator.services.mozilla.com/D225063
