merge_pr_54797

Use NewObject for Blob methods that return new instances

This change updates the IDL to match the W3C spec by explicitly marking
methods that always return a new object with NewObject.

This CL has no behavior changes.

Bug: 436420911
Change-Id: I42d3c048f3dd1edae2ae4fd4f791fd127d807d8e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6824263
Reviewed-by: Fergal Daly [email protected]
Reviewed-by: Jinho Bang [email protected]
Commit-Queue: Jinho Bang [email protected]
Cr-Commit-Position: refs/heads/main@{#1513751}

merge_pr_54796

[CodeHealth] Refactor pointerevent_pointerrawupdate.https.html

This is a no-op CL to improve the WPT for secure contexts:

• Replaces the state-machine based assertions with a simple event
  order assertion, which also adds the missing coverage for "one
  pointerrawupdate per pointermove" (as per
  https://w3c.github.io/pointerevents/#the-pointerrawupdate-event ).
• Adds the coverage for global event listeners.
• Ports old-style tests to promise_tests and "use strict".

(Confirmed manually that Firefox passes the test before and after.)

Bug: 404479704
Change-Id: I8dfb473b87872424fdb992ef5bc4ed0587949f45
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6936027
Auto-Submit: Mustaq Ahmed [email protected]
Reviewed-by: Kevin Ellis [email protected]
Commit-Queue: Kevin Ellis [email protected]
Cr-Commit-Position: refs/heads/main@{#1513732}

merge_pr_54794

Don't try to sort complex-typed CSS math operations based on its unit

When simplifying math expressions for sum or product we should normally
sort each component of the expression based on its type, e.g. 1rem + 1px

• 1em should become 1em + 1px + 1rem. However, when we have a
  complex-typed expression (e.g. 1rem / 1px * 1px) the spec is not clear
  about which type should be used for sorting. This leads to the crash in
  our code when we call ResolvedUnitTypeForSimplification, as that
  function doesn't know how to get a type of a typed expression.

So, this CL treats such complex-typed expressions as other expressions
that can't be sorted based on its type (such as css math functions, e.g.
sign()), meaning that sorting first puts numeric values sorted by their
alphabetical order of units and then it puts every other expression in
the same order it appeared in the original expression.

Also, this CL moves calls to ResolvedUnitTypeForSimplification close to
where it will be used (with numeric literals with double value), so that
for complex-typed expressions we don't call that function.

The test case also has 1px * 1px / 1px which should be simplified to
just 1px, but that is out of the scope of this CL.

Fixed: 443107131
Change-Id: I4078cd9d644b517d0ed0147beb2214118436750f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6919380
Reviewed-by: Steinar H Gunderson [email protected]
Commit-Queue: Daniil Sakhapov [email protected]
Cr-Commit-Position: refs/heads/main@{#1513663}

merge_pr_54791

Sync interfaces/ with @webref/idl 3.67.1 (#54791)

Co-authored-by: wpt-pr-bot [email protected]

merge_pr_54790

Revert "Bug 1720570: apply code formatting via Lando" for causing build bustages @ ModuleLoaderBase.h

This reverts commit ba4e61647555a5460cf544102b89d5b0363ff3e7.

Revert "Bug 1720570 - Update web-platform test expectations; r=layout-reviewers,emilio"

This reverts commit 27214f9ae8620d03eb481d7b0599f5176250e4fb.

Revert "Bug 1720570 - Use backgroundImage in relative-urls.html; r=layout-reviewers,emilio"

This reverts commit 97ec45d1dbceb0907db36c7e9a29f498012d9a1d.

Revert "Bug 1720570 - Support fetching css module scripts; r=dom-core,smaug"

This reverts commit e54b57ec16ba1e103bf664752bf77c74a0445b3b.

Revert "Bug 1720570 - Add support for css modules in ModuleLoader, WorkerModuleLoader and WorkletModuleLoader; r=allstarschh,dom-core,emilio,smaug"

This reverts commit 2eb55b783d07e484daeb49d159c42115621122be.

Revert "Bug 1720570 - Break out separate method to create constructed stylesheets; r=layout-reviewers,firefox-style-system-reviewers,emilio"

This reverts commit 2ccebf50bb7671ba540ea53ad5391ce9e21c760b.

Revert "Bug 1720570 - Move ModuleTypeAllowed check into ModuleLoaderBase; r=allstarschh"

This reverts commit 040b387d3e6af5bbeb4076d5377f989f3e14eb0c.

Revert "Bug 1720570 - Handle css module in GetModuleType; r=allstarschh"

This reverts commit 4b64b94cb100d525a21f1fe838561f084b5b636a.

Revert "Bug 1720570 - Add CreateCssModule; r=allstarschh"

This reverts commit 3f20b1d5b74c033864bb38e46932ded6485ea910.

Revert "Bug 1720570 - Add ModuleType::CSS; r=allstarschh"

This reverts commit ef5e3a71514a5e65e355334b0dc2a39df2ed267b.

Revert "Bug 1720570 - Add pref for css module scripts; r=layout-reviewers,emilio"

This reverts commit 7a9b1fbac0622f541985f3fc8843a0f37c83efd7.

gecko-commit: fe97e3b0105335b19a1df665a286b9ba3c28c124

merge_pr_54789

Adjust the font size of font-size-adjust-009.html to make Windows happy

The test font-size-adjust-009.html assumes that the Noto font has a
cap-height of 1462/2048 units. On Linux and macOS, the reported font
metrics match this assumption. However, Windows (DirectWrite) reports a
cap-height of 1472/2048, a 10-unit difference that causes the test to
fail.

This change adjusts the test font size so that the discrepancy is
diluted when floating-point values are truncated, ensuring consistent
results across platforms.

Bug: 441537622
Change-Id: I8d72717cf4643363b7c931c2b6e1b789d075e79f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6933761
Reviewed-by: Dominik Röttsches [email protected]
Commit-Queue: Dominik Röttsches [email protected]
Cr-Commit-Position: refs/heads/main@{#1513597}

merge_pr_54788

Use document.fonts.ready.then to not depend on font timing (#54788)

merge_pr_54773

layout: Set baseline even if line box has no fragment

InlineFormattingContextLayout::finish_current_line_and_reset() has an
early return in case the line has no fragment. However, if the line only
has a forced line break, then we still need to set the baseline.

Signed-off-by: Oriol Brufau [email protected]

merge_pr_54781

Stop including credentials on for registeringOrigins .well-known fetches

This isn't especially harmful, given that this is a security protection
rather than a privacy one, but .well-known fetches should generally be
made without credentials.

Change-Id: I59b191403deda4a558a6dac8116c6d73fafbfa23
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6923659
Reviewed-by: thefrog [email protected]
Commit-Queue: Daniel Rubery [email protected]
Cr-Commit-Position: refs/heads/main@{#1513247}

merge_pr_54780

ContainerTiming: do not use onload for setting up the WPT tests

After running the tests without onload, and as the observer is expected to
track the actual paints, it seems it is not really needed for set up. So this
changeset removes the usage of window.onload. This was inherited from the
elementtiming tests.

Bug: 382422286
Change-Id: Ibdd78cb84fb4f9778f5444115570cc86a3e8d393
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6596752
Commit-Queue: José Dapena Paz [email protected]
Reviewed-by: Michal Mocny [email protected]
Cr-Commit-Position: refs/heads/main@{#1513197}