Bug 52061 – [chromium] Transitions-related layout tests are failing more

• Status: RESOLVED
• Resolution: WONTFIX
• Component: Layout and Rendering

[Stephen White]

The following layout tests were flaky before, now failing more consistently post-r75169:

transitions/suspend-transform-transition.html = IMAGE PASS
transitions/change-values-during-transition.html = PASS TEXT
transitions/interrupted-accelerated-transition.html = PASS TEXT
transitions/transition-timing-function.html = PASS TEXT

Note that this is only in debug.  Perhaps timing-related?

[James Robinson]

I'm planning to revert r75169 (https://bugs.webkit.org/show_bug.cgi?id=51952#c15) which should help things.  That test was consistently failing for me in debug in the apple port DumpRenderTree on SnowLeopard pret-75169, though.

[Dean Jackson]

I have some updates coming in this area too.

[Dean Jackson]

Is this still happening after http://trac.webkit.org/changeset/81613 + http://trac.webkit.org/changeset/81722 ?

[James Robinson]

It might be - transitions/interrupted-accelerated-transition.html has failed twice on windows debug after 81722 landed.  The other bots+tests haven't cycled enough to make a good guess as to whether they are still affected.

http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=transitions%2Fsuspend-transform-transition.html%2Ctransitions%2Fchange-values-during-transition.html%2Ctransitions%2Finterrupted-accelerated-transition.html%2Ctransitions%2Ftransition-timing-function.html&group=%40ToT%20-%20chromium.org has the test history for those four tests on ToT WebKit + ToT Chromium bots (warning: very information-dense display).

[Dean Jackson]

bummer :( I'll take a look at the dense information. Thanks.

[Dean Jackson]

I think transitions/transition-timing-function.html is simply a slow bot. The results are nearly right - the test is just firing a little late.

[Dean Jackson]

Same with transitions/interrupted-accelerated-transition.html. The value is close, and the animation is happening.

Note that when 54151 lands for real, we'll be able to easily jump around in animations which should avoid this issue.

[Dean Jackson]

This leaves transitions/suspend-transform-transition.html as the test that seems to be clearly failing (although again, it *could* be a timing thing).

[Dean Jackson]

And that's also a timing issue. I think the best thing is to block this on 54151, then remove all timing dependent code from animation/transition tests.

[Stephen Chenney]

Test related bugs being marked WontFix. TestExpectations will contain bug if still relevant.