The following tests are failing / timing out on Sierra WK1 bots after "Use asynchronous ResourceHandleClient calls for WebKit1" <https://trac.webkit.org/changeset/223271/webkit> Regressions: Unexpected text-only failures (8) http/tests/loading/basic-auth-resend-wrong-credentials.html [ Failure ] http/tests/media/video-redirect.html [ Failure ] http/tests/security/contentSecurityPolicy/audio-redirect-allowed.html [ Failure ] http/tests/security/contentSecurityPolicy/video-redirect-allowed.html [ Failure ] http/tests/security/contentSecurityPolicy/video-with-http-url-allowed-by-csp-media-src-star.html [ Failure ] http/tests/security/cross-frame-access-put.html [ Failure ] http/tests/security/local-video-source-from-remote.html [ Failure ] http/tests/security/video-cross-origin-accesssameorigin.html [ Failure ] Regressions: Unexpected timeouts (8) http/tests/canvas/webgl/origin-clean-conformance.html [ Timeout ] http/tests/media/autoplay-if-audio-is-playing.html [ Timeout ] http/tests/media/video-load-twice.html [ Timeout ] http/tests/media/video-play-progress.html [ Timeout ] http/tests/security/canvas-remote-read-remote-video-localhost.html [ Timeout ] http/tests/security/canvas-remote-read-remote-video-redirect.html [ Timeout ] http/tests/security/video-cross-origin-caching.html [ Timeout ] http/tests/xmlhttprequest/xmlhttprequest-overridemimetype-invalidstaterror.html [ Timeout ] https://build.webkit.org/results/Apple%20Sierra%20Release%20WK1%20(Tests)/r223271%20(5274)/results.html
Based on the changelog in r223271, tests may be less deterministic than they used to be, so I don't know if this is a definitive list of the newly-flaky / failing tests.
I can only reproduce http/tests/loading/basic-auth-resend-wrong-credentials.html rarely and http/tests/xmlhttprequest/xmlhttprequest-overridemimetype-invalidstaterror.html
I think we should just skip whatever tests start being flaky on that patch, since they will be WK1 changes only which are intentional, especially if they are only on certain operating systems.
Also seems to affect these tests: http/tests/loading/basic-auth-resend-wrong-credentials.html http/tests/loading/basic-credentials-sent-automatically.html https://build.webkit.org/results/Apple%20El%20Capitan%20Debug%20WK1%20(Tests)/r223301%20(4232)/results.html I'll update expectations if that is the route we really want to take, but I still feel uneasy about it. Alexey, what are your thoughts?
I don't think that it's OK to skip these tests on WK1. Please roll back until a better solution can be found.
Rolled out in https://trac.webkit.org/changeset/223408. Duping to original bug. *** This bug has been marked as a duplicate of bug 160677 ***
http/tests/loading/basic-credentials-sent-automatically.html is being fixed in https://bugs.webkit.org/show_bug.cgi?id=178365
http/tests/xmlhttprequest/xmlhttprequest-overridemimetype-invalidstaterror.html was fixed in https://bugs.webkit.org/show_bug.cgi?id=150095 http/tests/loading/basic-auth-resend-wrong-credentials.html was fixed in https://bugs.webkit.org/show_bug.cgi?id=178370 http/tests/security/cross-frame-access-put.html was fixed in https://bugs.webkit.org/show_bug.cgi?id=178371
Since I cannot reproduce any of the remaining tests, I'm going to mark them as flaky and try speculative fixes until the bots are happy.