Bug 254745

Summary:	REGRESSION (261912@main): [macOS WK2] imported/w3c/web-platform-tests/css/css-sizing/aspect-ratio/replaced-element-035.html is a flaky failure
Product:	WebKit	Reporter:	Ryan Haddad <ryanhaddad>
Component:	Media	Assignee:	Antoine Quint <graouts>
Status:	REOPENED ---
Severity:	Normal	CC:	graouts, heycam, webkit-bot-watchers-bugzilla, webkit-bug-importer
Priority:	P2	Keywords:	InRadar
Version:	Other
Hardware:	Unspecified
OS:	Unspecified
See Also:	https://bugs.webkit.org/show_bug.cgi?id=241376 https://bugs.webkit.org/show_bug.cgi?id=257075

Description Ryan Haddad 2023-03-30 09:20:35 PDT

imported/w3c/web-platform-tests/css/css-sizing/aspect-ratio/replaced-element-035.html is a flaky failure on macOS WK2 queues. It is known to be flaky on macOS WK1, but the diff is not the same.

Image diffs showing that the test expects 00:00 but is seeing -00:00
https://build.webkit.org/results/Apple-Ventura-Debug-AppleSilicon-WK2-Tests/261918@main%20(2051)/imported/w3c/web-platform-tests/css/css-sizing/aspect-ratio/replaced-element-035-diffs.html

https://results.webkit.org/?suite=layout-tests&test=imported%2Fw3c%2Fweb-platform-tests%2Fcss%2Fcss-sizing%2Faspect-ratio%2Freplaced-element-035.html

Comment 1 Ryan Haddad 2023-03-30 09:23:20 PDT

Though there were some intermittent failures before, the flakiness on macOS WK2 is very apparent as of 3/21/2023.

cc'ing Antoine as there was a media controls related change at that time (commits.webkit.org/261912@main), but it isn't obvious to me that it is related.

Comment 2 Radar WebKit Bug Importer 2023-03-30 09:23:30 PDT

<rdar://problem/107423870>

Comment 3 Ryan Haddad 2023-03-30 13:04:32 PDT

Pull request: https://github.com/WebKit/WebKit/pull/12195

Comment 4 EWS 2023-03-30 13:10:07 PDT

Test gardening commit 262359@main (0157e68dea08): <https://commits.webkit.org/262359@main>

Reviewed commits have been landed. Closing PR #12195 and removing active labels.

Comment 5 Antoine Quint 2023-04-02 05:03:44 PDT

Ryan is right: this regressed with 261912@main.

Comment 6 Antoine Quint 2023-04-02 13:21:25 PDT

We end up in a layout state where we display the remaining duration time label instead of the remaining time label and this is due to some odd layout code in TimeControl. I believe the issue is that the metrics computed with _performIdealLayout() relies on the visibility state of the elapsed time label, I'm looking into cleaning things up there.

Comment 7 Antoine Quint 2023-04-02 13:23:51 PDT

Also worth looking into the flakiness with this test with WK1 (bug 241376).

Comment 8 Antoine Quint 2023-04-11 03:26:10 PDT

Pull request: https://github.com/WebKit/WebKit/pull/12610

Comment 9 EWS 2023-04-11 11:07:20 PDT

Committed 262829@main (ad969c4574b6): <https://commits.webkit.org/262829@main>

Reviewed commits have been landed. Closing PR #12610 and removing active labels.

Comment 10 Cameron McCormack (:heycam) 2023-05-20 23:49:51 PDT

Reverted in https://bugs.webkit.org/show_bug.cgi?id=257075 for causing the scrubber to break.