We fixed bug #234940 but did not add a regression test.
Created attachment 449412 [details] Patch
Comment on attachment 449412 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=449412&action=review > LayoutTests/media/video-src-blob-perf.html:19 > + failTest("Loading a data URL takes too long time."); It might be helpful to log the multiple when the test fails, to help diagnose failures in case the test is flaky in some situations. > LayoutTests/media/video-src-blob-perf.html:43 > + video1.src = url; > + start1 = Date.now(); It might be more accurate to set the source of both elements after the data url has been created. > LayoutTests/media/video-src-blob-perf.html:60 > + setTimeout(() => failTest("Cannot complete the test in 5 seconds."), 5000); Even normal tests sometimes take more than five seconds on a heavily loaded bot, so I would increase the timeout significantly (10 sec?)
Comment on attachment 449412 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=449412&action=review >> LayoutTests/media/video-src-blob-perf.html:19 >> + failTest("Loading a data URL takes too long time."); > > It might be helpful to log the multiple when the test fails, to help diagnose failures in case the test is flaky in some situations. Good point! Will fix it. >> LayoutTests/media/video-src-blob-perf.html:43 >> + start1 = Date.now(); > > It might be more accurate to set the source of both elements after the data url has been created. Agree. Actually, after this change, the multiple is about 2 in my local tests. >> LayoutTests/media/video-src-blob-perf.html:60 >> + setTimeout(() => failTest("Cannot complete the test in 5 seconds."), 5000); > > Even normal tests sometimes take more than five seconds on a heavily loaded bot, so I would increase the timeout significantly (10 sec?) Good point!
Created attachment 449418 [details] [fast-cq] Patch for landing
Committed r288221 (246181@main): <https://commits.webkit.org/246181@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 449418 [details].
<rdar://problem/87784063>