| Summary: | REGRESSION(257041@main): [ macOS ] 2X transitions/ (Layout-Tests are constant timeouts | ||
|---|---|---|---|
| Product: | WebKit | Reporter: | Robert Jenner <jenner> |
| Component: | CSS | Assignee: | Oriol Brufau <obrufau> |
| Status: | RESOLVED FIXED | ||
| Severity: | Normal | CC: | obrufau, simon.fraser, webkit-bot-watchers-bugzilla, webkit-bug-importer |
| Priority: | P2 | Keywords: | InRadar |
| Version: | WebKit Nightly Build | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
|
Description
Robert Jenner
2022-11-28 17:35:52 PST
I was able to reproduce the timeouts at Ventura Debug ToT running the test as follows: run-webkit-tests <insert_test_here> Test gardening commit 257098@main (e9b769950822): <https://commits.webkit.org/257098@main> Reviewed commits have been landed. Closing PR #6902 and removing active labels. I tried running the tests with the --no-timeout flag as well. After 10 minutes each the tests never ran. So these do appear to be legit timeouts, and not just slow tests. So I've added [ Timeout ] to the test's expectations at the commit above. It should also be noted that both of these tests were failing before 257041@main. There is just an expectation marked as [ Pass Failure ] Starting on Oriol who committed the change that caused these two tests to constantly timeout. Ah, that's because I made getComputedStyle().getPropertyCSSValue() consistently return a CSSValuePair, which throws when transition-test-helpers.js attempts to use getFloatValue(). The solution seems to use getPropertyValue() instead. Also adding a try..catch seems a good idea. Pull request: https://github.com/WebKit/WebKit/pull/6915 Committed 257208@main (4ee897dd423c): <https://commits.webkit.org/257208@main> Reviewed commits have been landed. Closing PR #6915 and removing active labels. There's equivalent code in LayoutTests/animations/animation-test-helpers.js (sadly) |