The media/media-blocked-by-willsendrequest.html test has been flaky across all chromium platforms for as far back as we have history. My guess is that the flakiness isn't limited to chromium, though. When the test fails, we are missing the following two lines of output: EVENT(loadstart) EVENT(error) From the symptoms, my best guess is that the test is trying to run before the video's loadstart. Eric, assigning to you just because you are the test author and I don't want to leave this ownerless. Please feel free re-title if this is chromium specific and/or bounce back to me for any reason. Chromium flakiness dashboard results: http://test-results.appspot.com/dashboards/flakiness_dashboard.html#showExpectations=true&tests=media%2Fmedia-blocked-by-willsendrequest.html
Created attachment 115154 [details] Patch
Committed r100275: <http://trac.webkit.org/changeset/100275>
(In reply to comment #0) > The media/media-blocked-by-willsendrequest.html test has been flaky across all chromium platforms for as far back as we have history. My guess is that the flakiness isn't limited to chromium, though. When the test fails, we are missing the following two lines of output: > EVENT(loadstart) > EVENT(error) > > From the symptoms, my best guess is that the test is trying to run before the video's loadstart. > I think your guess is correct. I believe that the test passes on other platforms but fails on chromium because the chromium media engine takes longer to start up so the widow's 'load' event is fired while the media events are in the queue waiting to be dispatched. > Eric, assigning to you just because you are the test author and I don't want to leave this ownerless. > I will change the test to not finish until all required events have been handled.
Created attachment 115165 [details] Proposed patch
Comment on attachment 115165 [details] Proposed patch Thanks! Any chance I could talk you into reverting r100275 along with this patch?
(In reply to comment #5) > (From update of attachment 115165 [details]) > Thanks! Any chance I could talk you into reverting r100275 along with this patch? Good point, will do.
http://trac.webkit.org/changeset/100282