Bug 1734939 [wpt PR 31169] - Update wpt metadata, a=testonly wpt-pr: 31169 wpt-type: metadata

Bug 1734939 [wpt PR 31169] - [@layer] Change layer order from preorder to postorder, a=testonly Automatic update from web-platform-tests [@layer] Change layer order from preorder to postorder This patch follows from a recent CSSWG resolution [1] that puts the default layer at the highest priority, and a subsequent spec edit [2] that changed the layer ordering from the preorder of the layer tree to the postorder. Most test changes just reflect this spec change. In some WPT tests, rule ordering is swapping so that the winning rule will appear first, so that we can verify layer is working. external/wpt/css/css-cascade/presentational-hints-cascade.html is moved from the stable test suite to a new disable-custom-element-default-style virtual test suite because this is the flag that actually makes the test result different on test and stable builds. We need to disable this flag while keeping cascade layers enabled to test the new test case in the file. [1] https://github.com/w3c/csswg-drafts/issues/6284#issuecomment-937262197 [2] https://github.com/w3c/csswg-drafts/commit/a6fc16e1b65868c6f984918cc76ad9b238e7139e Fixed: 1257936 Change-Id: I899de8c8a921b0535296957ee1b0cca1f60a8c61 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3214603 Reviewed-by: Anders Hartvoll Ruud <andruud@chromium.org> Reviewed-by: Xianzhu Wang <wangxianzhu@chromium.org> Commit-Queue: Xiaocheng Hu <xiaochengh@chromium.org> Cr-Commit-Position: refs/heads/main@{#930689} -- wpt-commits: fd7a9e9a139e01cd5017b57b8d3123d3a9a1a7bb wpt-pr: 31169

Bug 1735373 [wpt PR 31206] - Update wpt metadata, a=testonly wpt-pr: 31206 wpt-type: metadata

Bug 1735373 [wpt PR 31206] - Put paint offset for OOF in fragmentainer in ContainingBlockContext, a=testonly Automatic update from web-platform-tests Put paint offset for OOF in fragmentainer in ContainingBlockContext We compensate PaintOffsetTranslations for OOF paint offset in fragmentainer by subtracting, but we need to track the paint offset separately for different containing block contexts, otherwise we may subtract the offset of a PaintOffsetTranslation that is not in the ancestor transform chain of the OOF. For example, <div style="columns: 2; height: 40px"> <div style="height: 50px"> <div style="position: relative"> <div id="non-containing-block" style="will-change: opacity"> <div id="abs" style="position: absolute"></div> </div> </div> </div> "non-containing-block" creates a PaintOffsetTranslation that doesn't apply on "abs", but previously the offset was subtracted from the "abs". Also simplify the paint offset tracking. The original code was like delta = current_paint_offset - fragmentainer_paint_offset + adjustment_for_oof_in_fragmentainer; current_paint_offset -= delta; which is equivalent to current_paint_offset = fragmentainer_paint_offset - adjustment_for_oof_in_fragmentainer; So can just store one field paint_offset_for_oof_in_fragmentainer. Its initial value is the fragmentainer's paint offset and it is subtracted by the offset when we create a PaintOffsetTranslation. The new test passes with CompositeAfterPaint and LayoutNGBlockFragmentation because 1. CompositeAfterPaint fragments composited layers correctly (for "non-containing-block") 2. LayoutNGBlockFragmentation handles out-of-flow in multicol correctly. Change-Id: I03dcacf52f5810b91e49c6f8bf1d4fe4879e9a81 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3218477 Commit-Queue: Xianzhu Wang <wangxianzhu@chromium.org> Reviewed-by: Morten Stenshorne <mstensho@chromium.org> Cr-Commit-Position: refs/heads/main@{#930658} -- wpt-commits: 6ea1fa9437b658dceb6e8150084db4a2bc046b16 wpt-pr: 31206

Bug 1722519 [wpt PR 29804] - Add support for rerunning testharness(+testdriver) tests, a=testonly Automatic update from web-platform-tests Add support for rerunning testharness(+testdriver) tests (#29804) When running a test that uses testdriver using reload doesn't work, because the testdriver parts won't rerun. This is annoying. To make the development experience better, add a rerun button to testharness tests. In the case where the test is being run directly from the server, this just ends up calling window.reload(). In the case we're running via wptrunner, it causes the test to be rerun from the start. Implementation wise, this works by dispatching an event (currently named `__test_restart`) from the rerun button. This can be caught by the wait() implementation in executors, which now provide a return value of True to rerun the test and False to continue with the next test. There is an edge case where the manual reruns are mixed with automatic reruns specified on the command line. In practice it's not clear what the use case is, but the behaviour as written is to ensure the test is run at least as many times as specified on the command line, counting manual reruns against that total. -- wpt-commits: 335403228bc8359b6d74c75b4a5ac2efb0c767ce wpt-pr: 29804

Bug 1735310 [wpt PR 31202] - Don't assume the exit_after_start kwarg exists, a=testonly Automatic update from web-platform-tests Don't assume the exit_after_start kwarg exists In particular the tests don't pass this in. Defaulting to something falsey is reasonable. -- wpt-commits: 13ab4e86236d0ac82bb7f97e2bb7cf066ae0effb wpt-pr: 31202

Bug 1735336 [wpt PR 31204] - Don't let lines affect SubtreeModifiedMarginStrut-ness., a=testonly Automatic update from web-platform-tests Don't let lines affect SubtreeModifiedMarginStrut-ness. Checking the child margin when the child is an inline node doesn't make sense, since we'd then use the margin of the containing block node. We'd get a DCHECK failure for the SubtreeModifiedMarginStrut flag not being consistent between regular-ish [1] and simplified layout. In the testcase, there's a (default) BODY margin that would cause the problem. Initially, we'd set SubtreeModifiedMarginStrut of BODY to true (because there's inline content inside), then we'll trigger a relayout, by changing the width of an element on the last line inside BODY. In this pass we'll get a NGLayoutCacheStatus::kCanReuseLines hit, which will *not* set SubtreeModifiedMarginStrut. Then we'll trigger another relayout by modifying an OOF. Because it's an OOF change, the simplified layout machinery will be involved, and we'll CheckSameForSimplifiedLayout(). But in this pass SubtreeModifiedMarginStrut would be true again. [1] With a NGLayoutCacheStatus::kCanReuseLines hit. Bug: 1255718 Change-Id: Ib0bb1b56fbf41067f40544c89c026bf121cc7500 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3218972 Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org> Reviewed-by: Koji Ishii <kojii@chromium.org> Commit-Queue: Ian Kilpatrick <ikilpatrick@chromium.org> Cr-Commit-Position: refs/heads/main@{#930572} -- wpt-commits: 1cc2d8c2d9c42d26075f33d1cf8c43afea86152a wpt-pr: 31204

Bug 1735371 [wpt PR 30999] - Update wpt metadata, a=testonly wpt-pr: 30999 wpt-type: metadata

Bug 1735371 [wpt PR 30999] - [GridFragmentation] Allow grid container fragmentation., a=testonly Automatic update from web-platform-tests [GridFragmentation] Allow grid container fragmentation. This sets up the call to FinishFragmentation to support fragmenting grid containers. Grid layout is slightly unique as we work out the "grid" before performing the "final" layout on any of the children. As such we have the *total* intrinsic block-size, rather than the block-size up to the breakpoint. We'll likely need to iterate on this calculation. Bug: 614667 Change-Id: I3a43438175a42657599038c94677d995ea6285cf -- wpt-commits: 5609d1f13d9c2c6d5795e293f2a08aef3e0a0301 wpt-pr: 30999

Bug 1735357 [wpt PR 31205] - DOM: ensure event.path is not supported, a=testonly Automatic update from web-platform-tests DOM: ensure event.path is not supported -- wpt-commits: de8528fc713323073345bc76208350226c2d750f wpt-pr: 31205

Bug 1735242 [wpt PR 31195] - Bump pycodestyle from 2.7.0 to 2.8.0 in /tools, a=testonly Automatic update from web-platform-tests Bump pycodestyle from 2.7.0 to 2.8.0 in /tools Bumps [pycodestyle](https://github.com/PyCQA/pycodestyle) from 2.7.0 to 2.8.0. - [Release notes](https://github.com/PyCQA/pycodestyle/releases) - [Changelog](https://github.com/PyCQA/pycodestyle/blob/main/CHANGES.txt) - [Commits](https://github.com/PyCQA/pycodestyle/compare/2.7.0...2.8.0) --- updated-dependencies: - dependency-name: pycodestyle dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com> -- wpt-commits: 7a11185ba72730bcca63cc7b1bb1942a5f094228 wpt-pr: 31195

Bug 1735252 [wpt PR 31197] - Regenerate WPT certificates, a=testonly Automatic update from web-platform-tests Automated regeneration of WPT certificates -- wpt-commits: 2812b9b0e5830b56b1f2a4e00838c0a16b8db004 wpt-pr: 31197

Bug 1734072 [wpt PR 31109] - Skip conditional requirement install when `--skip-venv-setup` is passed, a=testonly Automatic update from web-platform-tests Skip conditional requirement install when `--skip-venv-setup` is passed -- wpt-commits: 299c886b69d9488d0e30552a35b702b122c63b70 wpt-pr: 31109

Bug 1735172 [wpt PR 31187] - Update wpt metadata, a=testonly wpt-pr: 31187 wpt-type: metadata

Bug 1735172 [wpt PR 31187] - [navigation-timing]: Test fine-grained Service Worker timing, a=testonly Automatic update from web-platform-tests [navigation-timing]: Test fine-grained Service Worker timing (#31187) * [navigation-timing]: Test fine-grained Service Worker timing Mark the timing inside the service worker (and adjust by timeOrigin), to make sure the timing reported to Navigation Timing correspond with the milestones specified in https://w3c.github.io/ServiceWorker/#service-worker-timing. The four relevant milestones which should be in order: - PerformanceNavigationTiming.workerStart - Finish the activation event - PerformanceNavigationTiming.fetchStart - Handling the Fetch event in the worker This is to ensure that the difference between workerStart and fetchStart measures the delay caused by service worker activation. See https://github.com/w3c/navigation-timing/issues/148 * Use step_timeout * Remove unnecessary timeouts -- wpt-commits: 2f8b6be9749e160f9902c9b2f99b00cf7b74b238 wpt-pr: 31187

Bug 1733749 [wpt PR 31077] - Update wpt metadata, a=testonly wpt-pr: 31077 wpt-type: metadata

Bug 1733749 [wpt PR 31077] - [css-fonts] Remove the "none" value from font-palette, a=testonly Automatic update from web-platform-tests [css-fonts] Remove the "none" value from font-palette (#31077) This value was removed in https://github.com/w3c/csswg-drafts/commit/b7103b0269cde51914af02f5ca1e5582d44e7b70. -- wpt-commits: 319699eab9ce7520797885e947604a8dde647867 wpt-pr: 31077

Bug 1734907 [wpt PR 31166] - [GridNG] Fix auto-placement with items spanning multiple rows (again), a=testonly Automatic update from web-platform-tests [GridNG] Fix auto-placement with items spanning multiple rows (again) Fixing an issue in the auto-placement algorithm where, if the current position moved past the `EndOnPreviousMajorLine` of an overlapping item, when calling `MoveToNextMajorLine` the major end line of such item would not be taken into account because it will be removed from the vector by the time its called. A similar issue was fixed previously, but at the time were items are introduced to the overlapping item's heap. Running a perf job showed that, as expected, no performance regression is introduced since this CL only changes some conditionals performed at the time of inserting/deleting any overlapping item. A new test to cover the regression was also added to the WPT suite. Bug: 1255298 Change-Id: Id5a5db3e0b999832767cd512794c5eebdff17286 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3213920 Commit-Queue: Ethan Jimenez <ethavar@microsoft.com> Reviewed-by: Ian Kilpatrick <ikilpatrick@chromium.org> Reviewed-by: Ana Sollano Kim <ansollan@microsoft.com> Cr-Commit-Position: refs/heads/main@{#930306} -- wpt-commits: 44642faf7d0f2156084c4d0689343786002bcb1c wpt-pr: 31166

Bug 1735089 [wpt PR 31181] - Update wpt metadata, a=testonly wpt-pr: 31181 wpt-type: metadata

Bug 1735089 [wpt PR 31181] - [WPT] Fix CSS response URL in fetch-request-css-base-url.https.html, a=testonly Automatic update from web-platform-tests [WPT] Fix CSS response URL in fetch-request-css-base-url.https.html The test expects that the referrer is the response URL of the CSS `fetch-request-css-base-url-style.css?fetch`, but '?fetch' is not added to the actual `fetch()` request for that CSS. After this CL the test passes on Chromium and Firefox. Bug: 915140 Change-Id: I637de67f4c3765b6cf90e8067022487a38f3eba2 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/3216924 Reviewed-by: Ben Kelly <wanderview@chromium.org> Commit-Queue: Ben Kelly <wanderview@chromium.org> Cr-Commit-Position: refs/heads/main@{#930156} -- wpt-commits: 8430b93260895f0bcd0dab03de168725cf4e7945 wpt-pr: 31181