Chromium Main Console

bluetooth: Clean up active notify sessions on ForgetDevice

WebBluetoothServiceImpl::ForgetDevice() revoked device access on the
legacy permissions backend by removing the device address from
allowed_devices(). However, it did not tear down active GATT notify
sessions, pending connection requests, or watch advertisement clients
associated with the forgotten device. If the physical connection was
kept open by another frame or origin, BLE characteristic notification
events continued to be dispatched to the renderer.

This CL introduces CleanupDeviceState() and invokes it during
ForgetDevice() when running on the legacy permissions backend. It closes
the GATT connection, cancels pending connection attempts, removes watch
advertisement clients, and drains in-flight and deferred notify session
requests with GATT_NOT_AUTHORIZED.

Bug: 495752713
Change-Id: Ia15abe31452f31f9ac2a7b0b6affc510d145122f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8494652
Reviewed-by: Reilly Grant <reillyg@chromium.org>
Commit-Queue: Rob Pitkin <robpitkin@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710628}

[soft navs] Remove kSoftNavigationTraceEvents killswitch

This has been in Stable since M145.

Bug: 465706684
Change-Id: Ied59c86cd81adcf94d48f1d7b93cf3040c9ff07b
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8499512
Commit-Queue: Johannes Henkel <johannes@chromium.org>
Auto-Submit: Scott Haseley <shaseley@chromium.org>
Reviewed-by: Johannes Henkel <johannes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710626}

Settings Lit: Remove Polymer version of RelaunchMixin.

Bug: 393471368
Change-Id: I3c5500694b1afae7878480424edb07e6c0c9af17
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8504263
Reviewed-by: Rebekah Potter <rbpotter@chromium.org>
Auto-Submit: Demetrios Papadopoulos <dpapad@chromium.org>
Commit-Queue: Rebekah Potter <rbpotter@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710625}

WebUI: Disallow conditional DOM nodes in $ interface, part 2.

In this part fixing all remaining violations of [1] where conditional
DOM nodes are declared as part of CrLitElement's helper '$' interface.

This is in preparation of adding an automated check to catch these.

[1]
https://chromium.googlesource.com/chromium/src/+/HEAD/docs/webui/webui_lit_style_guide.md#Static-Elements

Bug: 567169220
Change-Id: Ib9e8fc6e2a9b58f2fa1fae9b4ffff191b4457d4c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8508865
Auto-Submit: Demetrios Papadopoulos <dpapad@chromium.org>
Commit-Queue: Rebekah Potter <rbpotter@chromium.org>
Reviewed-by: Rebekah Potter <rbpotter@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710624}

Expect failure for positional-methods-with-parser-options.html

In crrev.com/c/8494796 (refs/heads/main@{#1710333}), the WPT test
external/wpt/domparsing/tentative/positional-methods-with-parser-options.html
was updated to expect createParserOptions(options, type, sink) to receive
the type and sink arguments.

However, passing type and sink to createParserOptions is pending
implementation in in-flight crrev.com/c/8496967. Consequently, the test
consistently fails on 10 headless_shell_wpt_tests bots (Linux Tests,
WebKit Win10, linux-bfcache-rel, linux-blink-leak-rel, mac13/14/15/26-arm64)
with:
assert_equals: expected (string) "TrustedHTMLParserOptions" but got (undefined) undefined

Mark the test as [ Failure ] until crrev.com/c/8496967 lands.

Bug: 491743369
Change-Id: I854139056edc0af5c23560f6d43761cbadc99bc2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8509245
Reviewed-by: Steven Luong <stluong@chromium.org>
Commit-Queue: Darryl James <dljames@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710622}

Prefer staged Android-synced theme in CrossDeviceSettingImporter

Before maybeApplyBackgroundUpdateFromDeviceSync() runs, an
Android-synced theme is staged in getSyncedNtpBackgroundData() while
getNtpBackgroundData() holds stale or null state. Without preferring the
staged theme, cross-OS imports diff against stale state and Undo reverts
to the old/default theme instead of the user's Android-synced theme.

Bug: 517615321
Change-Id: I08adcb1964bccc9144150bb7511d1adb9601a1a7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8430682
Reviewed-by: Xinyi Ji <xinyiji@chromium.org>
Commit-Queue: Jenna Himawan <jhimawan@google.com>
Reviewed-by: Mark Schillaci <mschillaci@google.com>
Cr-Commit-Position: refs/heads/main@{#1710621}

Add unit tests to client.openscreen.chromium builders

Configure the downstream Open Screen Chromium builders
(chromium_linux_x64, chromium_mac_arm64, chromium_win_x64) on the
client.openscreen.chromium waterfall to run media_unittests,
cast_unittests, and mirroring_unittests.

Bug: 567669801, 568834969
Change-Id: I5b74a30996f9df1331e8421692d6f0f0a4fa6837
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8487190
Commit-Queue: Jordan Bayles <jophba@chromium.org>
Reviewed-by: Ben Pastene <bpastene@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710617}

[glic] Fix Android mic permission flow not showing

With GlicNoWebview, GlicPwcPermissionDelegate decides Gemini's mic
permission, which skips Chrome's usual prompt for a missing Android
RECORD_AUDIO permission. Tapping the mic then failed silently: with the
Gemini mic setting off it was denied, and with the setting on but the
Android permission revoked, capture failed with no prompt or snackbar.

To make GlicNoWebview behave like the webview, which always allows mic
requests on all platforms:
- GlicPwcPermissionDelegate always grants MEDIASTREAM_MIC. This also
applies to desktop GlicNoWebview, matching desktop webview.
- After the user accepts Chrome's mic dialog, the Android permission is
requested directly instead of through MediaCaptureDevicesDispatcher.
If it is denied, the mic disabled snackbar is shown. If the activity
is destroyed before Android answers, the request is denied.
- The Android mic permission UI is owned by GlicKeyedService so tests
can replace it.

Adds an Android browser test for the flow with and without
GlicNoWebview.

Follow-ups: b/568819844 (Gemini mic setting doesn't gate mic access),
b/568819419 (same issue for location), b/568870385 (reuse Chrome's
standard Android permission reprompt instead of a Glic-specific one).

TAG=agy
CONV=83375f7a-30f4-49a1-92ae-9acc4ebee8ba

Bug: 568192203
Change-Id: I468e6e45034dc3e2d8db6aa9b9949751d56c0ad5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8490187
Reviewed-by: Hailey Wang <haileywang@google.com>
Reviewed-by: Dan H <harringtond@chromium.org>
Commit-Queue: Siddhartha S <ssid@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710616}

[base/i18n] Bind DateTimeFormatter instances to a locale

DateTimeFormatter now receives the locale in its constructor and uses it
to format when no explicit locale is passed to Format().

IcuBridge owns one formatter per locale, cached in an
absl::flat_hash_map keyed by LanguageTag, and exposes:
- date_time_formatter(): returns the formatter for the current default
ICU locale, obtained from GetDefaultIcuLocale().
- date_time_formatter(locale): returns the formatter for the given
locale.

Formatters are created on first use. The map is lazily populated by the
const accessors and guarded by a lock, as IcuBridge is a process-wide
singleton that can be accessed from any thread.

This change is being done so a cache can be more conveniently
implemented for improving the performance of DateTimeFormatter.

Change-Id: I02d50c783b5c66e87ff24ab7e42cb21f819e8d1e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8454223
Reviewed-by: Yann Dago <ydago@chromium.org>
Reviewed-by: Friedrich Hauser <friedrichh@chromium.org>
Commit-Queue: Danilo Tedeschi <danft@google.com>
Reviewed-by: Ireneusz Szulc <ireneuszszulc@google.com>
Reviewed-by: Max Curran <curranmax@chromium.org>
Reviewed-by: Alexei Svitkine <asvitkine@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710615}

[iOS][AuroraAIM] Add Fieldtrial Config for AimButtonRefactor

Fixed: 568691840
Change-Id: Ic3b9e7adfeb103ee17beeab8d5ec6e25efb7ee18
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8503102
Commit-Queue: Eric Ekey <ericekey@google.com>
Reviewed-by: Chris Lu <thegreenfrog@google.com>
Cr-Commit-Position: refs/heads/main@{#1710614}

[blinkpy] Remove redundant default_expectations_files() overrides

AndroidPort and WebviewPort overrides default_expectations_files(). As
crrev.com/c/6520154 removed MobileTestExpectations, both overrides
return the same list as Port.default_expectations_files(), so this
removes them.

All callers iterate the result once or wrap it in list()/set(), so
returning the base class's filter object instead of a list makes no
difference.

It also removes the unused get_blink_dir import from android.py.

No behavior change.

Bug: 568602794
Change-Id: I16ffbbe59d9c5ad9d3b4c656910a160f7a6e42fe
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8507194
Reviewed-by: Jonathan Lee <jonathanjlee@google.com>
Commit-Queue: Julie Jeongeun Kim <jkim@igalia.com>
Cr-Commit-Position: refs/heads/main@{#1710613}

[iOS][AuroraAIM] Hide Standalone AIM when Ineligible

This change hides the standalone AIM module button beside the Most
Visited collection when AIM is not allowed on the NTP (such as when the
default search engine is not Google). When hidden, the MVT container
layout is updated for its new width.

Bug: 565801056, 549012340
Change-Id: I373f8a16b4d4a0f8541ba0e0872a1a1fd0ff2abf
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8503156
Commit-Queue: Eric Ekey <ericekey@google.com>
Reviewed-by: Chris Lu <thegreenfrog@google.com>
Cr-Commit-Position: refs/heads/main@{#1710612}

[Reading Mode] Don't turn unlinked anchors into links on toggle

Avoid setting data-link or href when a link has an empty URL, and strip
href from unmatched Readability anchors before converting them to spans
so toggling links off and back on does not turn them into links.

This is already done when building the tree the first time when links
are enabled. This CL just ensures it's done when toggling off and on
again.

Bug: 562113981
Change-Id: I717b0797e3826288cc92fd4e0e07cd9d84f11e6a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8499973
Commit-Queue: Kristi Saney <kristislee@google.com>
Reviewed-by: Lauren Winston <lwinston@google.com>
Auto-Submit: Kristi Saney <kristislee@google.com>
Cr-Commit-Position: refs/heads/main@{#1710610}

[Glic] Add location denial and policy disabled tests to GlicSettingsUnitTest

Add unit tests verifying that denying the OS location permission prompt
reverts GLIC_GEOLOCATION_ENABLED to false, and that enterprise policy
disabled state (isDisabledByPolicy) disables and unchecks settings
switches.

Bug: None
Change-Id: Ic99ad7c5ba4f89ffe0e74e30fad0f7c09883b178
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8494081
Commit-Queue: Hailey Wang <haileywang@google.com>
Reviewed-by: Mei Liang <meiliang@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710608}

Cache Storage: use BigBuffer instead of Blob for writing side data

The side data is generated in full in the renderer and held completely
in memory. Using Blob infrastructure for it is unnecessary.

Bug: 567175133
Change-Id: I7303092572aa4d28a2dd98f8e77c7f25e6e4bb94
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8483291
Reviewed-by: Eriko Kurimoto <elkurin@chromium.org>
Commit-Queue: Evan Stade <evanstade@microsoft.com>
Reviewed-by: Nasko Oskov <nasko@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710607}

remove xiangdongkong@ from OWNERS

I no longer work on Chromium

BUG: n/a
Change-Id: I2ed31a1c22e9dfdc4efa45329c34ba2595bb2714
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8509244
Auto-Submit: Xiangdong Kong <xiangdongkong@google.com>
Reviewed-by: Jimmy Gong <jimmyxgong@chromium.org>
Commit-Queue: Jimmy Gong <jimmyxgong@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710606}

[Test] Use long poll timeout for FirstRunIntegrationTest native init

launchFirstRunActivityAndWaitForNativeInitialization waits for native
initialization with the default 3s CriteriaHelper timeout. The class is
@DoNotBatch, so every method cold-starts native in a fresh process,
which occasionally exceeds 3s on loaded desktop emulators and causes CQ
shard retries. Use DEFAULT_MAX_TIME_TO_POLL_LONG instead.

Bug: 568844205
Change-Id: Ic38198f428ac1a8e493ec8de02c6fd17db9cd7e9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8508294
Reviewed-by: Jenna Himawan <jhimawan@google.com>
Commit-Queue: Sky Malice <skym@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710603}

[NTP Customization Sync] Plumb the attribution action URL to sync

When a theme collection image is applied by rebuilding the background
preference, i.e. when it is re-selected from the local theme history or
imported from another device, Android wrote an empty attribution action
URL and pushed that to sync. Desktop then lost the link on the
attribution, and iOS applied the synced background without it.

Carry the URL along with the two attribution lines:
- CustomBackgroundInfo gets a non-null attributionActionUrl, which is an
empty GURL when there is no link. Like the attribution lines it
describes the image rather than identifies it, so it is not part of
equals().
- The JNI factories, the cross-device theme tracker and the theme
history (from CollectionImage.attributionUrl) fill it in.
- onThemeCommitted() passes it down to
UpdateThemeCollectionPrefsWithColor(), which writes it instead of an
empty GURL.
- The theme history JSON stores it under "attributionActionUrl" when it
is not empty.

Bug: 488439751
TAG=agy
CONV=9a0411f4-0500-48f3-8472-93db77df2b0c

Change-Id: Ic32821b5e2d7f2010b508c326597a5941d0dd9ab
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8502200
Reviewed-by: Tibor Goldschwendt <tiborg@chromium.org>
Commit-Queue: Xinyi Ji <xinyiji@chromium.org>
Reviewed-by: Xi Han <hanxi@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710602}

Geolocation: Cache permission denied error in OnClientsChanged()

Ensure GeolocationProviderImpl::OnClientsChanged() populates the cached
permission denied error result (result_) when system_permission_status_
is kDenied.

Previously, OnClientsChanged() returned early without populating
result_ when system_permission_status_ was kDenied. If
GeolocationProviderImpl was lazily instantiated while system permission
was already denied (or after all previous clients disconnected and
cleared result_), AddLocationUpdateCallback() did not invoke the newly
registered callback with an error, leaving GeolocationImpl's
QueryNextPosition() request pending indefinitely.

Bug: 559426542
Change-Id: I2e502e69b378eeb1e3294f395f29be34c3b647b8
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8488527
Reviewed-by: Rob Pitkin <robpitkin@chromium.org>
Commit-Queue: Alvin Ji <alvinji@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710600}

Fix @SuppressWarnings("DoNotMock") in a few tests

Simple fixes for:
* MessageBannerCoordinatorUnitTest.java
* HomeButtonCoordinatorTest.java

Merged all test methods into one (since it now does more work):
* BottomSheetViewBinderUnitTest.java

Bug: 567604165
Change-Id: If3bb3bf8c9159327104ce7d0146465c426c1380f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8484463
Owners-Override: Henrique Nakashima <hnakashima@chromium.org>
Reviewed-by: Henrique Nakashima <hnakashima@chromium.org>
Commit-Queue: Andrew Grieve <agrieve@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710599}

[Glic] Keep Actor overlay visible behind the omnibox popover on Desktop

Hiding the overlay while the omnibox is focused is only needed when the
suggestions list is drawn below it. On Desktop, the suggestions are
drawn as an elevated popover above the overlay, so keep the overlay
visible behind it. The take-over button is still hidden while the
omnibox is focused.

Phones and tablets are unchanged: the overlay is still hidden while the
omnibox is focused.

Desktop:
http://screencast/cast/NjM2NDQ0NDgyNDE3NDU5Mnw0Y2E0NjgyZi05NQ#t=2

Mobile: http://screencast/cast/NjI2NzcwMDk1NDI2NzY0OHxmMjU5NTIwMS1jNw#t=2

TAG=agy

Fixed: 568104361
Change-Id: I21319c108bec8b675b6cc2dce9a223510ee6c423
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8489036
Commit-Queue: Linda Wang <lindaywang@google.com>
Reviewed-by: Wenyu Fu <wenyufu@chromium.org>
Reviewed-by: Hailey Wang <haileywang@google.com>
Cr-Commit-Position: refs/heads/main@{#1710597}

Settings Lit: Migrate remaining autofill_page/ to PrefService, part 3

In this part removing all remaining use of prefs from autofill_page
in favor of PrefService.

Bug: 393471368
Change-Id: I3565c1f52ab0c9382c5fee1e726e379e53dd8c11
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8478462
Reviewed-by: Demetrios Papadopoulos <dpapad@chromium.org>
Commit-Queue: Rebekah Potter <rbpotter@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710596}

[Crubit] Parallelize Crubit actions via metadata groups

Previously, `cc_bindings_from_rs` (Crubit) actions had a build-time
dependency on `_rspgen_target_name` (`generated_file`), which in turn
depended on `_original_attributes.deps` (e.g. `//build/rust/std:std_bindings`).
Because `std_bindings` and its constituent targets depend on upstream
Crubit header generation actions, this caused GN to emit Ninja hard
dependencies serializing all Crubit actions (e.g. `signed_web_bundles`
behind `std_bindings`, `std_bindings` behind `alloc_bindings`).

`cc_bindings_from_rs` only needs the GN metadata (header path strings,
namespaces, and features) written into the `.rsp` file at `gn gen` time.
It does not consume the generated C++ header files at runtime.

This CL introduces a pure metadata target `_metadata_target_name`
(`${target_name}_crubit_metadata`) including `std_bindings_crubit_metadata`
that contains only GN metadata without any action or library dependencies.
`_rspgen_target_name` depends on these metadata targets instead of the
action-bearing targets, and `_action_target_name` explicitly lists
`_action_rsp_path` in its `inputs` and depends on `_rspgen_target_name`.

This:
1. Breaks the Ninja harddeps chain between Crubit actions, allowing
all Crubit actions to run fully concurrently as soon as their Rust
libraries build.
2. Ensures `_action_rsp_path` is explicitly declared in `inputs` so
remote execution (RBE) workers receive the `.rsp` file.
3. Preserves `_original_attributes.deps` on `source_set(_h_target_name)`
so that downstream C++ compilations still wait for generated headers.

Bug: 568640261, 563098656
Change-Id: I4eff80137b9c28d15b04b8a1f8f7500acd806372
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8505621
Commit-Queue: Takuto Ikuta <tikuta@chromium.org>
Auto-Submit: Takuto Ikuta <tikuta@chromium.org>
Commit-Queue: Łukasz Anforowicz <lukasza@chromium.org>
Reviewed-by: Łukasz Anforowicz <lukasza@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710594}

[Vertical Tabs] Rename VerticalTabHoverController to VerticalTabItemHoverController.

Pure renaming, no logic/behavior changes.

Bug: 542280452
Change-Id: I1f5c93a25f7b4e418731046afa519a5a9494efef
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8508451
Reviewed-by: Gazal Agarwal <agazal@google.com>
Commit-Queue: Hao Dong <spdonghao@google.com>
Cr-Commit-Position: refs/heads/main@{#1710593}

[contextual-tasks] Add interactive UI test for Ask Google context menu entrypoint.

Tests the default side panel flow, the omnibox flow, and the submenu
side panel flow.

Bug: 567601617
Change-Id: I205a355b06b49d0f36bc3864155b009ea51eadef
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8504784
Reviewed-by: Duncan Mercer <mercerd@google.com>
Commit-Queue: Bryan Nguyen <nguyenbryan@google.com>
Cr-Commit-Position: refs/heads/main@{#1710592}

Check IsUsingVulkan in IsDefaultANGLEVulkan

Ensure that IsDefaultANGLEVulkan() immediately returns false if Vulkan
is not being used.

Fixed: 566777649
Change-Id: Ic1a05e462173732afe0e1b1a02bbaa424e0ef896
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8496482
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710590}

Fix AccountPreviewPreference usage in account aware password promo test

In crrev.com/c/8460583 (refs/heads/main@{#1710062}),
account_aware_password_signin_promo_unittest.cc was added using
pref.other_device_form_factor. Shortly after, in crrev.com/c/8500215
(refs/heads/main@{#1710150}), other_device_form_factor was combined into
PreferredDeviceInfo other_device_info.

Because crrev.com/c/8500215 branched before #1710062, it missed updating
the newly added test file. This causes compile failures on
Chrome-branded builders such as chrome/ci/win-chrome:

../../chrome/browser/ui/webui/password_manager/notification_cards/account_aware_password_signin_promo_unittest.cc(98,10):
error: no member named 'other_device_form_factor' in
'signin::AccountPreviewDataService::AccountPreviewPreference'
98 | pref.other_device_form_factor = form_factor;
| ~~~~ ^

Update the test to set pref.other_device_info.form_factor and include
syncer::PASSWORDS in enabled_data_types so device-aware titles format
correctly.

Bug: 546059799, 567901894
Change-Id: Ia4eea043c041eafb62025e107073be83bdf410fb
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8499848
Reviewed-by: Steven Luong <stluong@chromium.org>
Commit-Queue: Darryl James <dljames@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710589}

blink/dom: Add RAW_PTR_EXCLUSION and STACK_ALLOCATED annotations

Covers core/dom.

Marks stack-allocated cache scope pointers in Document and static table
fields in EventTarget with RAW_PTR_EXCLUSION.

Also converts QualifiedNameComponents and QualifiedNameData from
DISALLOW_NEW() to STACK_ALLOCATED() because they are only instantiated
on the stack (as temporary lookup keys for AddWithTranslator and
HashComponents; QualifiedNameCache stores QualifiedNameImpl* rather than
QualifiedNameData, and HashComponents requires unique object
representations via base::byte_span_from_ref).

Bug: 566006164
Change-Id: I7d930229a9cbcb165136b99a38fe68a46efb2789
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8457725
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Commit-Queue: Keishi Hattori <keishi@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710587}

Remove CSS ruby tests from content-shell.filter

The following tests run on headless shell and can be removed from
content_shell.filter: external/wpt/css/css-ruby/block-ruby-003.html
external/wpt/css/css-ruby/ruby-inlinize-recursive-simple.html
external/wpt/css/css-ruby/ruby-text-combine-upright-001a.html
external/wpt/css/css-ruby/ruby-text-combine-upright-001b.html

One test fails on Mac, but still runs on headless shell, so add it to
TestExpectations to better reflect its status:
[ Mac ] external/wpt/css/css-ruby/ruby-inlinize-recursive-simple.html

Bug: 526667990
Change-Id: I2696c80d4eae752582ab3fe437e0d7f5df222bf0
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8503258
Reviewed-by: Philip Rogers <pdr@chromium.org>
Commit-Queue: Sharon Yang <yangsharon@igalia.com>
Cr-Commit-Position: refs/heads/main@{#1710586}

Side Panel: Prepare for auto formatting

In this CL, left aligning templates in .html.ts files and removing
unnecessary clang-format on/off comments, to make the formatting diff
easier to review.

Bug: 512911634
Change-Id: I7d3705db6c0b84a7a01a1d302f510d39e3675cd5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8477554
Auto-Submit: Rebekah Potter <rbpotter@chromium.org>
Commit-Queue: Rebekah Potter <rbpotter@chromium.org>
Reviewed-by: Demetrios Papadopoulos <dpapad@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710585}

blink/trustedtypes: Add RAW_PTR_EXCLUSION to stack-scoped table structs

Covers core/trustedtypes.

Marks stack-scoped table struct fields in BuildAttributeVector() and
BuildPropertyVector() with RAW_PTR_EXCLUSION.

Fields updated: 3 fields marked RAW_PTR_EXCLUSION in 1 file.

Bug: 566006164
Change-Id: Ief4a861643e9e1186c12328f576cc63397595a1d
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8467243
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Commit-Queue: Keishi Hattori <keishi@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710584}

Add missing widevine cdm headers dependency to //chrome/common:unit_tests

In crrev.com/c/8504759 (refs/heads/main@{#1710425}),
cdm_registration_unittest.cc was moved from //chrome/test:unit_tests
into //chrome/common:unit_tests. However, only data_deps was added,
omitting the direct header dependency.

This caused GN check to fail during generate_build_files on the
chrome/ci/linux-cft builder where enable_widevine is true:

The target:
//chrome/common:unit_tests
is including a file from the target:
//third_party/widevine/cdm:headers

Suggestion: Under if (enable_widevine && is_linux), add deps = [
"//third_party/widevine/cdm:headers" ] to :unit_tests (defined at
//chrome/common/BUILD.gn:869)

Add the missing dependency to resolve the build failure.

Bug: 521960574
Change-Id: I7e841f3f27cf9352a7aa6414786c97fc5d0a4fde
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8508977
Reviewed-by: Steven Luong <stluong@chromium.org>
Commit-Queue: Darryl James <dljames@chromium.org>
Reviewed-by: Foromo Daniel Soromou <koretadaniel@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710583}

Restore TestExpectations for src-then-lazy-load.html

Test is still timing out, add it back to the test expectations list.

Change-Id: Ic01a8cadda1ff0f7522dc3b34ec55a0c9bee3bc8
Bug: 375209498
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8508568
Reviewed-by: Foromo Daniel Soromou <koretadaniel@chromium.org>
Commit-Queue: Darryl James <dljames@chromium.org>
Reviewed-by: Steven Luong <stluong@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710582}

Identify ComposeboxMenuSharedTab by input item identifier

Replace `serverToken` with `inputItemIdentifier` on
`ComposeboxMenuSharedTab` and its removal delegate/mediator methods
(`didRemoveTabWithInputItemIdentifier:` and
`removeSharedTabWithInputItemIdentifier:`).

`ComposeboxMenuSharedTab` is a UI view model for the Shared Tabs menu in
Cobrowse. It only needs to identify the corresponding
`ComposeboxInputItem` in `ComposeboxInputPlateMediator` when the user
removes a shared tab. Using the local `ComposeboxInputItem.identifier`
(which exists immediately upon item creation) instead of `serverToken`
(which is only populated once a server upload starts) decouples the UI
menu from the server upload lifecycle and prepares for lazily uploading
auto-added tabs.

Bug: 568219593
Change-Id: I726d996ab0fe73bb4c63cd9a71ca5a8fcaccb8ba
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8495620
Commit-Queue: Louis Romero <lpromero@google.com>
Reviewed-by: Ameur Hosni <ameurhosni@google.com>
Reviewed-by: Aliona Dangla <alionadangla@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710581}

blink/misc: Convert raw pointer fields to UnprotectedInRelease

Covers core/fileapi, core/frame, core/html, core/loader, core/paint,
core/sanitizer, core/style and core/timing.

These fields get no BRP protection in release builds (the trait resolves
to RawPtrNoOpImpl, matching the previous raw pointer's cost). This is a
first step to enable dangling pointer detection on fuzzers and tests,
and we plan to convert suitable fields to regular protected raw_ptr in
the future.

Also marks static table and stack-scoped fields with RAW_PTR_EXCLUSION,
and marks HTMLViewSourceDocument::Link as STACK_ALLOCATED() because it
is only instantiated on the stack.

Fields converted: 5 across 5 files (plus 13 fields marked
RAW_PTR_EXCLUSION across 9 files and 1 struct marked STACK_ALLOCATED()).

Bug: 566006164
Change-Id: I8fcd09bf288e9095ecf0cd40c0333f277161467e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/8457728
Reviewed-by: Jeremy Roman <jbroman@chromium.org>
Commit-Queue: Keishi Hattori <keishi@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1710579}