Closed nt1m closed 10 months ago
@danielsakhapov @emilio Can you review for Blink / Gecko?
I think it would make sense to defer style containment from Interop until we have definitions on how it affects other named things:
etc.
I would like the spec to at least progress to that point, with a model that makes sense.
As it stands now, I'm not sure including style containment is very useful without those defined.
Since the mentioned issue has been resolved, should we close this issue?
I think it's a bit of a late clarification to address this in Interop 2023 (and it seems like no browser currently implements everything correctly either). I would still be in favor of removing the tests in the context of 2023 and having this re-apply for next year.
I don't think we should remove the style containment tests. They were included in https://github.com/web-platform-tests/wpt-metadata/pull/3356 which is the added scope labeled interop-2023-contain
that we reviewed before launching Interop 2023.
Spec changes to clarify things discovered while implementing have happened throughout the year. Has there been a significant scope change here?
Based on the initial failing tests, @danielsakhapov has fixed quote and counter containment in Chrome, specifically for Interop 2023, so dropping that now doesn't seem quite right.
@nt1m I went ahead and added label:interop-2023-contain
do your search queries to exclude a few tests that aren't in Interop 2023 anyway.
Looking through changes to the tests that have happened this year, I find https://github.com/web-platform-tests/wpt/pull/41799, rewriting contain-style-counters-004.html. IIUC the old test was wrong, but since that didn't go through a test change proposal, removing the test is OK if that change on its own makes sense. (I presume that restoring the old test doesn't make sense, otherwise that should be the default outcome in case of disagreement.)
Test List
https://wpt.fyi/results/css/css-contain?label=master&label=experimental&aligned&q=contain-style%20label%3Ainterop-2023-contain https://wpt.fyi/results/css/css-contain?label=master&label=experimental&aligned&q=counter-scoping%20label%3Ainterop-2023-contain https://wpt.fyi/results/css/css-contain?label=master&label=experimental&aligned&q=quote%20label%3Ainterop-2023-contain
Rationale
https://github.com/w3c/csswg-drafts/issues/9212 is open for discussion. This needs clarification in the CSSWG. Also it's unclear to me whether the behavior that is tested is a sensible one, given the complexity required for the implementators. A sensible spec for something like style containment shouldn't be complex to implement.