nsContainerFrame.h was only using the enum nsLayoutUtils::IntrinsicISizeType,
which this patch moves to LayoutConstants.h instead.
Depends on D91505
Differential Revision: https://phabricator.services.mozilla.com/D91506
The only thing in nsIFrame.h that was using it was the implementation
of a templated method, which could be moved to nsIFrameInlines.h.
Depends on D91504
Differential Revision: https://phabricator.services.mozilla.com/D91505
This patch covers thw low-hanging fruit: headers that either don't use
nsLayoutUtils.h at all, or where only method implementations that can
easily be moved to the .cpp file use it.
Differential Revision: https://phabricator.services.mozilla.com/D91504
It stopped being relative to mScrollOffset in bug 1519285, when that
patched change the value stored in mScrollOffset to be the visual
scroll offset even for the Layers copy of the metrics.
Depends on D91479
Differential Revision: https://phabricator.services.mozilla.com/D91480
The two existing uses were inconsistent in whether the margins
were relative to the visual viewport or the layout viewport.
We could fix this, but it turns out both uses are readily
refactored to not need the field, so this patch removes the
field.
Differential Revision: https://phabricator.services.mozilla.com/D91479
Constructing a GCCellPtr means getting the trace kind from a lookup table based
on the alloc kind. We can move this out of line so that code's not included in
the barrier.
Differential Revision: https://phabricator.services.mozilla.com/D91312
This moves barrier implementations out of derived cell types, with the types
supplying necessary information to a central implementation (e.g. with the
isPermanentAndMayBeShared() method).
Differential Revision: https://phabricator.services.mozilla.com/D91310
Rather than calling GCMarker::traverseEdge directly, call do DoMarking because
this also handles checking/clearing the expected compartment information in
debug builds, the lack of which is causing this assertion failure.
Differential Revision: https://phabricator.services.mozilla.com/D91405
This change allows to retrieve `arguments[Symbol.iterator]` without calling the
resolve hook and without actually adding the iterator function to the arguments
object.
Differential Revision: https://phabricator.services.mozilla.com/D91031
Both arguments objects resolve hook were defining the iterator property, but
didn't set the corresponding `ITERATOR_OVERRIDDEN_BIT` flag. We don't need to
apply a similar change for the other resolved properties, because they're
using the `(Un)MappedArgGetter` and `(Un)MappedArgSetter` property functions.
In preparation for part two, the code to retrieve `ArrayValues` was already
moved into a separate function.
Differential Revision: https://phabricator.services.mozilla.com/D91030
This renames the internal -moz-math-script-level property in order to
prepare for full math-depth support. Currently, the property is guarded
under a disabled-by-default flag, so there should be no observable
behavior change.
Differential Revision: https://phabricator.services.mozilla.com/D91285
The last Avast Antivirus's hook function contains `CALL [disp32]` instruction.
Our detour needs to be able to handle that pattern.
Differential Revision: https://phabricator.services.mozilla.com/D91155
This patch optimizes our detour's code handling Opcode 0xFF, expanding
its coverage to INC and DEC reg64 as well as PUSH and CALL.
Testcases for these scenarios are of course included.
Differential Revision: https://phabricator.services.mozilla.com/D91154
- Rename EnsurePrinterInfo to TryEnsurePrinterInfo to reflect the fact that it can fail to populate the printer info.
- Add argument to TryEnsurePrinterInfo to accept a connection other than the default.
- Allow TryEnsurePrinterInfo to try once with the default connection and once with an established connection.
- Use an established connection to retrieve the paper list information.
Differential Revision: https://phabricator.services.mozilla.com/D91193
This patch calls setSearchMode directly from search(). This sets up a solution for the problem in the bug and also fixes the issue where a call to search() with a restriction token would flicker the token before it was replaced with the search mode indicator. I added new tabmenu and bookmarkmenu entry points to take advantage of this new functionality.
This also fixes the issues with handoff. Besides the problem of search() recording typed for handoff, `handoff` wasn't even registered as a Telemetry probe! That was my mistake. I added a test for handoff telemetry. It was only practical to test it in PBM since it uses a different implementation than about:home that's easier to test. I wrote a lengthy comment above the subtest about why I think this is okay.
Differential Revision: https://phabricator.services.mozilla.com/D91076
Automatic update from web-platform-tests
[wpt] Check ChromeDriver version before accepting found binary (#25657)
This fixes a long standing annoyance where we will blindly accept any
'chromedriver' binary, even if it isn't compatible with the Chrome
version under test. Instead, check that the versions of Chrome and
ChromeDriver mostly match. ChromeDriver is allowed to lag behind by a
release version, to allow for using a local build of Chromium.
See https://github.com/web-platform-tests/wpt/issues/10451
--
wpt-commits: 28bab3e887e6ea4e4398d8e069fd1f92421f1e95
wpt-pr: 25657
Automatic update from web-platform-tests
Refine crossOriginIsolated implementation for workers (#25748)
- Fix WindowOrWorketGlobalScope.crossOriginIsolated behavior for
workers.
- For SharedWorkers and ServiceWorkers it always return false. See
https://crbug.com/1131403 and https://crbug.com/1131404.
- Rename ExecutionContext::IsCrossOriginIsolated to
CrossOriginCapability. This is aligned with
https://html.spec.whatwg.org/C/#concept-settings-object-cross-origin-isolated-capability.
- Fix wpt/html/infrastructure/safe-passing-of-structured-data/shared-array-buffers/blob-data.https.html.
I originally planned to do that in
https://github.com/web-platform-tests/wpt/pull/24600 but I couldn't
do that due to some flakiness.
- Add more tests for workers in
wpt/html/cross-origin-embedder-policy/cross-origin-isolated-permission.https.html.
Bug: 1115379, 1018680, 1131403, 1131404
Change-Id: I2afcb01403f67a11fd06aefde1238aba16b68f36
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2416428
Commit-Queue: Yutaka Hirano <yhirano@chromium.org>
Reviewed-by: Domenic Denicola <domenic@chromium.org>
Reviewed-by: Robert Flack <flackr@chromium.org>
Reviewed-by: Hiroki Nakagawa <nhiroki@chromium.org>
Cr-Commit-Position: refs/heads/master@{#810130}
Co-authored-by: Yutaka Hirano <yhirano@chromium.org>
--
wpt-commits: ce281cc3f32d8e93ecbb33a51321d846ee2aae37
wpt-pr: 25748
Automatic update from web-platform-tests
Web Share: restrict URL scheme to http and https
We now follow the recent spec change limiting the permitted scheme
for shared urls to http and https - see
https://github.com/w3c/web-share/issues/173https://github.com/w3c/web-share/pull/174https://github.com/w3c/web-share/pull/177
We make an exception if the page performing the share it itself loaded
from a different scheme (e.g. file) - in that case we allow the same
scheme to be used for the shared url.
Bug: 1131755
Change-Id: I6abf0f9acd40ef79ec49379314e2ef3a81d3467e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2425977
Commit-Queue: Eric Willigers <ericwilligers@chromium.org>
Reviewed-by: Glen Robertson <glenrob@chromium.org>
Auto-Submit: Eric Willigers <ericwilligers@chromium.org>
Cr-Commit-Position: refs/heads/master@{#810180}
--
wpt-commits: a28408e23e7cb1e4e8dc070445a51fc2f2d9a4e6
wpt-pr: 25755
Automatic update from web-platform-tests
Increase the maximum number of open files for wpt's servers https://bugs.webkit.org/show_bug.cgi?id=215829
Patch by Sam Sneddon <gsnedders@apple.com> on 2020-09-21
Reviewed by Youenn Fablet.
macOS has a much lower limits than other OSes by default, and the iOS bots often run into it with their level of parallel test execution. This bumps the limit for each wptserve process up to 2048, which is double the limit on Debian, and should probably be safe.
* web-platform-tests/tools/serve/serve.py:
(ServerProc.create_daemon):
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@267354 268f45cc-cd09-0410-ab3c-d52691b4dbfc
--
REGRESSION(r267354): Only import resource module on Darwin https://bugs.webkit.org/show_bug.cgi?id=216823
Patch by Sam Sneddon <gsnedders@apple.com> on 2020-09-22
Reviewed by Darin Adler.
* web-platform-tests/tools/serve/serve.py:
Move the resource import to the local point behind the Darwin platform
test. This code is only run once per process initiation, hence it makes
no difference to have the import locally.
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@267417 268f45cc-cd09-0410-ab3c-d52691b4dbfc
--
wpt-commits: 64cdc2d8769e69f9b08bdff153e973a88aa015aa, b3b32e61d700d3bd9877c755e3aa943b0c468e0c
wpt-pr: 25773
Automatic update from web-platform-tests
Fix tests that return values to async_test in webrtc/ (#25719)
These tests are written correctly, they just need to not return the
promises they use as part of the flow.
They could be rewritten to use promise_tests if desired, but that would
have performance implications (as it changes them from concurrent to
sequential) so leaving that for now.
See https://github.com/web-platform-tests/wpt/issues/21435
--
wpt-commits: db9d339c441f03b2b89b00bce3662fb6e043c6fb
wpt-pr: 25719
