nsINavHistoryResultNode.guid doesn't exist, so this always falls back to uris;
unfortunately separators don't have a uri.
Differential Revision: https://phabricator.services.mozilla.com/D205716
Longer-term we could make `toSorted` itself a trampoline native, but that's more
complicated and this patch gets us most of the performance and code reuse benefits.
Differential Revision: https://phabricator.services.mozilla.com/D204629
This ports Array.prototype.sort to C++ code with a JIT trampoline. The sort algorithm
is the same as the self-hosted code to avoid subtle differences in behavior.
Any data used by the sorting algorithm is stored in the `ArraySortData` instance stored
in the trampoline frame. The `sortWithComparator` C++ function is like a generator
that we can yield from and resume into. This avoids slow calls from C++ into JIT code.
Sort calls with a JS comparator function are much faster with this implementation.
Especially in Baseline, the implementation is more than 3x faster on micro-benchmarks.
Some Speedometer 3 subtests improve by 2-3%.
Differential Revision: https://phabricator.services.mozilla.com/D204627
This patch adds code for using the JitEntry mechanism for native builtins.
The TrampolineNative JSJitInfo mechanism is very similar to what we do for
InlinableNative. For each trampoline native, a JIT trampoline is generated and
we use this as the function's JitEntry when creating the native JSFunction.
Differential Revision: https://phabricator.services.mozilla.com/D204625
Simplify the handling of Ion frames by checking for Ion frames directly instead of
checking for all the other frame types and excluding those. This is less of a
footgun when adding new calls to this function.
Differential Revision: https://phabricator.services.mozilla.com/D204624
Automatic update from web-platform-tests
Apply auction config's reportingTimeout.
Both top level and component auction config can specify a
reportingTimeout field to control the timeout of their reportWin() and
reportResult(). If component auction does not provide one, it will use
top level config's reporting timeout. A reporting timeout longer than 5
seconds will be clamped to 5s. If no reporting timeout is provided, a
default timeout kScriptTimeout is applied, as before this change.
Feature detection of this will be in a separate CL.
Fixed: b/326922003
Change-Id: Ibe2b41310dc44e6834a32ffa14afab0dd6ecb687
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5374544
Commit-Queue: Qingxin Wu <qingxinwu@google.com>
Reviewed-by: Ken Buchanan <kenrb@chromium.org>
Reviewed-by: mmenke <mmenke@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1277143}
--
wpt-commits: aab5ae6d81b0bd2b2ad3a9553de9d3da2f2bf675
wpt-pr: 45238
Automatic update from web-platform-tests
Pass a Chrome flag to ignore FedCM well-known files
This is needed to make the tests work until the PAC support is fixed
for HTTPS URLs.
--
wpt-commits: f8e40089bf7d4d026248a3cbb5f0edaa27283b6a
wpt-pr: 45285
Automatic update from web-platform-tests
Implement media-progress() function
As introduced in https://drafts.csswg.org/css-values-5/#media-progress-func
The media-progress() functional notation returns a <number> value
representing current value of the specified media query as a progress value
between two explicit values of the media query.
Note: for now we only support width and height with manual resolution.
Once it's properly specced, proper MQ parser and evaluator should be
used.
Note: now it's an editor's draft with a very strong chances to be accepted.
Change-Id: I6bc2ae310d96ba45693e77aa76b15627b4cd66ce
Bug: 40944203
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5277337
Reviewed-by: Anders Hartvoll Ruud <andruud@chromium.org>
Commit-Queue: Daniil Sakhapov <sakhapov@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276992}
--
wpt-commits: 65aa1254fd5f8d9c4aa5c9961f039e7f4511eeab
wpt-pr: 44442
Automatic update from web-platform-tests
webnn: fix input validation tests for large dimensions
The existing test case didn't actually calculate the number of
elements because the input dimension value exceeded the max
`unsigned long`
Fixed: 330060043
Change-Id: Ie052139b28f86504b4476925d175f4c0c01c5b82
Cq-Include-Trybots: luci.chromium.try:win11-blink-rel
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5375859
Auto-Submit: Feng Dai <feng.dai@intel.com>
Reviewed-by: Austin Sullivan <asully@chromium.org>
Commit-Queue: Feng Dai <feng.dai@intel.com>
Reviewed-by: ningxin hu <ningxin.hu@intel.com>
Cr-Commit-Position: refs/heads/main@{#1276979}
--
wpt-commits: 4ab7ce11f813676c926d08dcc0e09cbda83402d3
wpt-pr: 45152
Automatic update from web-platform-tests
Test that imported module has initiator-type=script (#45276)
* Test that imported module has initiator-type=script
* Remove a `console.log` from file to appease linter
--
wpt-commits: f369f5f691df77be893029c9910b2901285da76b
wpt-pr: 45276
Automatic update from web-platform-tests
Apply an accessible name to form.
The spec says that the "form" role only makes it a "landmark" if there
is an accessible name.
ARIA 1.3 says that the form role is a sub-role of landmark:
https://w3c.github.io/aria/, so if there is no accessible role then
the "form" role is inappropriate.
Since HTML-AAM doesn't say what the role should be otherwise (maybe
it should be "section"? That's what Chromium currently does), I
did not add a testcase for that situation.
Bug: 41490744
Change-Id: I6c6a93485a55a6645547cdf9e4eb4c8751e45141
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5384597
Commit-Queue: Chris Harrelson <chrishtr@chromium.org>
Reviewed-by: Aaron Leventhal <aleventhal@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276910}
--
wpt-commits: 12366f9880df6c2d716b6c9ef75c32b5df0132ac
wpt-pr: 45265
Automatic update from web-platform-tests
Address slowness of WPT
prefer-targeted-element-main-frame.html is timing out on quite a few
builders[1]. On "Win10 Tests x64" for example, even when it passes, it
often takes ~5.5 seconds which is super close to the timeout limit of 6
seconds. This patch attempts to make the test less slow.
Prior to this patch, the test opened 9 windows, running one test per
window. With this patch, the test will run all 9 tests in the same
window by making use of the location.hash API.
If the slowness persists after this lands, we should probably just mark
it a SlowTest.
[1]https://ci.chromium.org/ui/test/chromium/ninja%3A%2F%2F%3Ablink_wpt_tests%2Fexternal%2Fwpt%2Fcss%2Fcss-scroll-snap%2Fsnap-after-relayout%2Fmultiple-aligned-targets%2Fprefer-targeted-element-main-frame.html
Bug: 330385828
Change-Id: I618163e2c716ed90333e8909127566c757cda217
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5384129
Commit-Queue: David Awogbemila <awogbemila@chromium.org>
Reviewed-by: Steve Kobes <skobes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276893}
--
wpt-commits: d5abd8103fd8eb365e933ccf1b4f104193e316d1
wpt-pr: 45254
Automatic update from web-platform-tests
Make window-open-blank-from-different-initiator.html use slow.py instead
The test initially used infinite-slow-response.py to make a navigation
that never commits. However, infinite-slow-response.py will actually
commit as the HTTP response for the navigation *started* (it just never
finished), unblocking the navigation from committing. To achieve the
goal of having a navigation that practically never commits, we will
use slow.py, which will not start the HTTP response until after a very
long delay, making a navigation that's pending for a very long time.
Bug: 330011153
Change-Id: I9a4778c71474f1f612a00f184d6abaad5b2612a9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5376394
Commit-Queue: Rakina Zata Amni <rakina@chromium.org>
Commit-Queue: Arthur Sonzogni <arthursonzogni@chromium.org>
Reviewed-by: Arthur Sonzogni <arthursonzogni@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276860}
--
wpt-commits: 95e7e99b22bdf6a1a688ccd76feb886cdf611cc8
wpt-pr: 45275
Automatic update from web-platform-tests
Fix non-hue channel keyword values for hsl() and hwb()
SL and WB are stored in blink::Color in range [0 1] range, while the
[0 100] range is expected. Adjust these parameters when setting up the
channel keywords. Make sure that all values - independent of type -
are adjusted in MakePerColorSpaceAdjustments() before writing into the
destination blink::Color.
Some testcases expected an alpha value of 1 to be serialized, and one
had a space too many. Fix those up.
Bug: 330096624
Change-Id: I3b3983d73153b035687a262dae9187afea626e6d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5383086
Commit-Queue: Fredrik Söderquist <fs@opera.com>
Reviewed-by: Aaron Krajeski <aaronhk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276791}
--
wpt-commits: f763dd7d7b7edb982d221cba45c4b13c550d1549
wpt-pr: 45253
Automatic update from web-platform-tests
Remove child content selector from StylableSelect UA rules
This patch removes the :-internal-select-author-datalist and
:-internal-select-author-button pseudoclasses. Based on the standards
discussions, it seems like we will always render the potentially
problematic select UA rules instead of trying to detect whether the new
content model is used or if the computed value of appearance is the new
value. The only problematic styles are the select's border and
background-color, which are probably not a big deal.
Bug: 1511354
Change-Id: I9261183e799c1df57ca2737acee6966137bafeae
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5367160
Reviewed-by: David Baron <dbaron@chromium.org>
Commit-Queue: Joey Arhar <jarhar@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276589}
--
wpt-commits: 0267c3951dca216ef608725676cb62853f37dfed
wpt-pr: 45222
Automatic update from web-platform-tests
Implement rules for when we can interpolate calc-size() expressions.
This starts interpolating between two calc-size() expressions, or
between calc-size() expressions and keywords. It supports most of the
cases where we should do so, though leaves some TODOS for calc-size()
nested within the basis.
The failing tests related to lack of support for the 'stretch' keyword.
Bug: 313072
Change-Id: Ie36719cad46649d5c1f449adbc6d4ad4dfc96308
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5374567
Reviewed-by: Robert Flack <flackr@chromium.org>
Commit-Queue: David Baron <dbaron@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276526}
--
wpt-commits: bf9b67634ea793234aff670093c6945bee92a614
wpt-pr: 45256
Automatic update from web-platform-tests
IDB: idbobjectstore_delete WPT extended to workers in addition to window
web_tests/external/wpt/IndexedDB/Tests currently only verify in window
environment. Extend tests to cover dedicated, shared and service
workers.
Tests related to idbobjectstore_delete.htm have been extended.
7 idbobjectstore_delete.htm files have been consolidated into one file:
idbobjectstore_delete.any.js along with
idbobjectstore_deleteindex.any.js
idbobjectstore_deleted.htm has been extended to
idbdatabase_deleteObjectStore.any.js.
The other existing files like idbdatabase_deleteObjectStore.htm (3-4)
have also been consolidated to idbdatabase_deleteObjectStore.any.js.
Bug: 41455766
Change-Id: Ied919a90027f30a86f8679dc463a0f7a2930d6bd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5371125
Commit-Queue: Sneha Agarwal <snehagarwal@microsoft.com>
Reviewed-by: Ayu Ishii <ayui@chromium.org>
Reviewed-by: Evan Stade <estade@chromium.org>
Reviewed-by: Rahul Singh <rahsin@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#1276500}
--
wpt-commits: 88149528a94fdf769c29e129f8979859e39c339c
wpt-pr: 45098
Automatic update from web-platform-tests
Implement mouse behavior for StylableSelect
This patch makes clicking the <option>s in the <select> choose the
option and close the popover.
Bug: 1511354
Fixed: 330534051
Change-Id: I3faca1eff5eda7c1c6162445e0d1f5e71e2a92f7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5352148
Reviewed-by: David Baron <dbaron@chromium.org>
Commit-Queue: Joey Arhar <jarhar@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276477}
--
wpt-commits: a885671ab12851f38db7c51f2a49c7cadd0c773e
wpt-pr: 44979
Automatic update from web-platform-tests
webnn: Remove long timeout for WPT validation_tests
These validation tests should all just hit synchronous exceptions in
the renderer, so they shouldn't need a long timeout
Change-Id: I1f50ac4a4a819d59a6f6954836dbf37e280553ff
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5380961
Reviewed-by: Joshua Bell <jsbell@chromium.org>
Commit-Queue: Joshua Bell <jsbell@chromium.org>
Auto-Submit: Austin Sullivan <asully@chromium.org>
Commit-Queue: Austin Sullivan <asully@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276428}
--
wpt-commits: e3869b9d52624b281acc3ba6b8739794134bd804
wpt-pr: 45260
Automatic update from web-platform-tests
[anchor] Support try-tactic flips for inset-area
Since inset-area is not yet fully supported within @position-try,
this CL tests the flipping behavior via computed value only.
Like the self-alignment properties, inset-area must always be present
in the try-tactics layer, because we need to transform the value
without switching the property itself.
Bug: 40279608
Change-Id: Ibd7bd786e256c2142590290e37e188806eadb7a3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5383432
Commit-Queue: Anders Hartvoll Ruud <andruud@chromium.org>
Reviewed-by: Rune Lillesveen <futhark@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1276424}
--
wpt-commits: 33125ab6b58409b8216e4bf97c774a90b09ddf95
wpt-pr: 45245
