The following layout test is flaky on GTK Linux 64-bit Release media/media-ended.html [ Timeout Pass ] The flakiness ins't linked to any particular commit: http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=media%2Fmedia-ended.html&showExpectations=true&showLargeExpectations=true
Created attachment 319461 [details] Patch
Comment on attachment 319461 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=319461&action=review > LayoutTests/media/media-ended.html:17 > - audio.src = findMediaFile("audio", "content/silence") > + audio.src = findMediaFile("audio", "content/short") good catch, but "short" is really short.. only 27 msecs! Not sure if is right to use such short audio file here. The main reason why this test was timing out on GTK bot, but not on other port bots was because our aggressive time out values. We raised that on r218270 (and is still lower than on Mac ports). I have checked with wktesthunter that this test is now passing always since r218270. So I think we can just remove the flaky expectation and close this bug (leaving the test as it was originally wrote with the silence audio file).
Created attachment 319585 [details] Patch
Comment on attachment 319585 [details] Patch Good catch Carlos!
Comment on attachment 319585 [details] Patch Clearing flags on attachment: 319585 Committed r221476: <http://trac.webkit.org/changeset/221476>
All reviewed patches have been landed. Closing bug.
<rdar://problem/34694188>