merge_pr_50862

Abstract away the types of reftest templates in canvas test generator

In many places, the test generator was working with the whole list of
possible reference template types. It's simpler to create a single list
enumerating these types and using that list everywhere. This will also
simplify adding new types of reference templates in the future.

Bug: 393561071
Change-Id: Id8a2211186eeb51439c7229b9821878950e9e825
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6219868
Commit-Queue: Jean-Philippe Gravel [email protected]
Reviewed-by: Andres Ricardo Perez [email protected]
Cr-Commit-Position: refs/heads/main@{#1423078}

merge_pr_50859

GetSourceImageForCanvas: Use kDoNotChangeAlpha in BaseRenderingContext2D

The calls to GetSourceImageForCanvas are in the functions drawImage,
createPattern, and drawMesh.

The conversion to premultiplied format for these functions is unnecessary work, because the downstream Skia calls (e.g,
SkCanvas::drawImage) will take into account the alpha format via the
SkAlphaType (assuming that the SkAlphaType is not dropped on the
ground).

Add a new WPT test to ensure that ImageBitmaps created as premultiplied
and unpremultiplied both work correctly. This test failed when using
OOP-raster when the cc::ImageTransferCache dropped the SkiaAlphaType
on the ground. This tests passes on FireFox, but fails one sub-case
on WebKit (likely due to unpremultiplying already unpremultiplied
ImageData when creating an ImageBitmap).

Bug: 4027721
Change-Id: Iede7051d18b96a0f502e2c24745119018313cfec
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6280927
Commit-Queue: ccameron chromium [email protected]
Reviewed-by: Colin Blundell [email protected]
Cr-Commit-Position: refs/heads/main@{#1423050}

merge_pr_50858

WebKit export of https://bugs.webkit.org/show_bug.cgi?id=284545 (#50858)

merge_pr_50857

Improve GetListItems descendant tracking This patch improves and consolidates some tree traversal code for invalidating HTMLSelectElement::GetListItems. , , and elements won't be included in any of these traversals if they are nested inside of an , an , or inside nested s. This also improves the OwnerSelectElement performance of HTMLOptGroupElement and HTMLHRElement by caching the closest select ancestor on insertion. Fixed: 396769395 Change-Id: Ib584f2914ee46b0206eebcb93ed9d41f0ae3796b Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6279485 Reviewed-by: Traian Captan [email protected] Commit-Queue: Joey Arhar [email protected] Cr-Commit-Position: refs/heads/main@{#1422956}

merge_pr_50856

Revert "Move Chromium implementation specific WPT."

This reverts commit e4ac5529356c17e412989c1fb335318910653314.
w3c/ServiceWorker#1752 standardized the
limitations. Let me revert the commit that moved such WPTs
to the Chromium specific directory.

Additional changes:

• Fixed conflicts.
• Updated the link from Chromium code to the ServiceWorker specification
  because it is now standardized.

Bug: 395384783

Change-Id: I0a057fdc2b26d76e14caad75e7f8e07951f1c6f3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6289645
Reviewed-by: Keita Suzuki [email protected]
Reviewed-by: Minoru Chikamune [email protected]
Commit-Queue: Yoshisato Yanagisawa [email protected]
Reviewed-by: Shunya Shishido [email protected]
Cr-Commit-Position: refs/heads/main@{#1422938}

merge_pr_50854

Rename gentestutilsunion.py to gentest.py

gentest.py was nothing more than a wrapper around gentestutilsunion.py.
This CL deletes the former and renames the latter (keeping its content
otherwise unchanged).

Bug: 40207206
Change-Id: I58efcad63267ec459a66db9b59e737c130c2fcd7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6207155
Reviewed-by: Andres Ricardo Perez [email protected]
Commit-Queue: Jean-Philippe Gravel [email protected]
Cr-Commit-Position: refs/heads/main@{#1422936}

merge_pr_50853

Rebase and squash

Signed-off-by: Xiaocheng Hu [email protected]

merge_pr_50852

Fix a crash induced by DocumentPartRoot::clone() of a detached doc

Bug: 384233287, 40276946
Change-Id: I5bd59c22fca80be55dba9fafb7ea3e22423cda6a
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6287265
Reviewed-by: Mason Freed [email protected]
Commit-Queue: Andrey Kosyakov [email protected]
Cr-Commit-Position: refs/heads/main@{#1422889}

merge_pr_50851

Convert css/filter-effects/filter-subregion-01.html to a real ref test

The original reference was a bitmap with a lot of differences from
real rendering, e.g. antialiased pixels along diagonal lines and
incorrect color of the filtered result. It also failed on Firefox.

The new reference uses real SVG like the test, with the subregion
filters replaced with full filters on rects to test the subregion
filter behavior.

Fixed: 41432730
Change-Id: I40d7aa310ebae7205f24db747b2e9a9702ad30aa
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6289325
Auto-Submit: Xianzhu Wang [email protected]
Commit-Queue: Philip Rogers [email protected]
Reviewed-by: Philip Rogers [email protected]
Cr-Commit-Position: refs/heads/main@{#1422861}

merge_pr_50838

WebKit export of https://bugs.webkit.org/show_bug.cgi?id=287789 (#50838)