Revision r191948 <http://trac.webkit.org/changeset/191948> has caused 4 new timeouts: Regressions: Unexpected timeouts (4) media/video-controls-no-scripting.html [ Timeout ] media/video-source-error.html [ Timeout ] media/video-source-load.html [ Timeout ] media/video-source-moved.html [ Timeout ] I have tested to build r191947 and r191948, and I can confirm that this 4 tests started to time out with r191948. Note: media/video-controls-no-scripting.html was failing before because of bug 141835 (but it was not timing out).
Carlos, do you have time to look at this?
No, but I'll find it.
Created attachment 270485 [details] Patch This seems to be a problem in the very few cases that the gst bus sync handler is called form the main thread. We were avoiding scheduling to the main thread in that case, but for some reason the bus gets stalled or something. So always end the message to the main thread. This fixed the timeouts for me for media/video-source-load.html, media/video-source-moved.html and media/video-controls-no-scripting.html, the others timeout, but they also timeout for me reverting the patch.
Comment on attachment 270485 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=270485&action=review > Source/WebCore/ChangeLog:9 > + sync hanlder was called in the main thread. It seems that typo: hanlder
Comment on attachment 270485 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=270485&action=review > Source/WebCore/platform/graphics/gstreamer/MediaPlayerPrivateGStreamer.cpp:1906 > + GRefPtr<GstMessage> protectMessage(message); protectedMessage is a better name for this local variable, since it’s not just used to protect the message, but to send it across threads as well
Committed r196060: <http://trac.webkit.org/changeset/196060>