merge_pr_49653

moveBefore: fail when moving across roots

This applies the recent spec conversation in
whatwg/dom#1307 (comment)
where the moved & parent node have to be in the same
shadow-including root, not only in the same document.

Bug: 40150299
Change-Id: Ic4daa6e863a8e96061ad9bebad2264202d87f722
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6088082
Commit-Queue: Noam Rosenthal [email protected]
Reviewed-by: Dominic Farolino [email protected]
Cr-Commit-Position: refs/heads/main@{#1395407}

merge_pr_49652

Fixes to external/wpt/svg/types/SVGElement.ownerSVGElement-01.html

Fix scoping of the tests variable. Fix the selector for the <title>
element (was targeting the HTML title).

Bug: 383379792
Change-Id: Ic4f7162aa43d700822dd1b2faf4b1686d8f9f8ca
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6087969
Auto-Submit: Fredrik Söderquist [email protected]
Commit-Queue: Rune Lillesveen [email protected]
Reviewed-by: Rune Lillesveen [email protected]
Cr-Commit-Position: refs/heads/main@{#1395365}

merge_pr_49651

Add multicol anchor positioning tests for vertical-rl.

There was no test coverage for this.

Change-Id: I6f4d35fd1b4c5037c525e73642da3150ed368240
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6084481
Commit-Queue: Morten Stenshorne [email protected]
Reviewed-by: Koji Ishii [email protected]
Cr-Commit-Position: refs/heads/main@{#1395331}

merge_pr_49649

Make virtualenv use sysconfig to find paths

Instead of having our own copies of what we believe paths should be,
or manually mutating what sysconfig returns for other variables, we
should just delegate all of this to sysconfig in the standard way.

This fixes #49545, fixing Python 3.12.8 support, where sysconfig
stopped overly caching certain variables, thus leading to oddities
after we've rewritten sys.exec_prefix and sys.prefix.

merge_pr_49648

Make dialog light dismiss and dialog.requestClose not need UA

See the conversation here:

openui/open-ui#1128 (comment)

Previously, both dialog light dismiss and dialog.requestClose() would be
subject to CloseWatcher's abuse prevention logic, which requires user
activation. But that doesn't make sense, as pointed out in the thread
above, since neither of those are "abusive" things - they're not
keeping a user from navigating back. So this CL adds a parameter
to CloseWatcher.requestClose() that allows callers to proceed with
requesting close even if there isn't sufficient user activation.
Note that this CL does not change the existing behavior for other
CloseWatchers - those will still require user activation for now.
See crbug.com/383593252 for that.

Bug: 376516550,383593252
Change-Id: Ibf5ccc835794aba2aa627ddec7caafa6c9c26e55
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6085724
Commit-Queue: Mason Freed [email protected]
Auto-Submit: Mason Freed [email protected]
Reviewed-by: David Baron [email protected]
Cr-Commit-Position: refs/heads/main@{#1395143}

merge_pr_49646

Update test for OpenUI resolution

This patch adds test cases corresponding to this OpenUI resolution:
openui/open-ui#1119 (comment)

It also cleans up the leftover queueMicrotask calls which are no longer
needed now that things are synchronous.

Change-Id: I1031d2899197e313dd9eb20a142868930b886a28
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6086797
Reviewed-by: Traian Captan [email protected]
Commit-Queue: Joey Arhar [email protected]
Cr-Commit-Position: refs/heads/main@{#1395174}

merge_pr_49641

Don't consider collapsed margins for in-flow line bounds if we're not a BFC.

So that our client{Width,Height} and scroll{Width,Height} calculations
make sense.

Differential Revision: https://phabricator.services.mozilla.com/D231629

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1936156
gecko-commit: a661bff10b1fea7e6d828efad298bbdbe2a8574c
gecko-reviewers: dshin

merge_pr_49638

Expand web-platform-test coverage to include subdocument fetch credentials

Some of the tests discussed in privacycg/storage-access#210

Differential Revision: https://phabricator.services.mozilla.com/D231720

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1936355
gecko-commit: 8db61a9fffaff233a8ff6648033afd8f67a52f63
gecko-reviewers: anti-tracking-reviewers, timhuang

merge_pr_48486

[wptrunner] Support testdriver.js in {ref,print-ref,crash}tests (#48486)

• [wptrunner] Add testdriver flag to non-testharness manifest items

This will let wptrunner gracefully skip testdriver tests if the executor
doesn't support them.

• [wptrunner] Support testdriver.js in {ref,print-ref,crash}tests

Non-testharness executors that don't support testdriver simply skip
those tests.

merge_pr_49644

Deflake wpt/fledge/tentative/server-response.https.window.html

... in the "fails because test keys are not configured case".
It looks like one of the real-life keys got a short ID of 0x12,
so the "key ID isn't what the test configures" failed about a fifth
of the time... so change around the test ID to avoid it.

Bug: 383379786,383413865,383403583,383401572,383399676
Bug: 383397963,383387527,383395709,383393294,383391660
Bug: 383383789,383377137,383374957,383367270,383381981
Bug: 383379241,383372392
Change-Id: Ic27b16de689d9c94a24b65b1e5be19e1ff839afa
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6088048
Commit-Queue: Maks Orlovich [email protected]
Reviewed-by: Russ Hamilton [email protected]
Cr-Commit-Position: refs/heads/main@{#1395006}