As of writing, the Media Capture and Stream specification [1] has some inconsistencies with regards to the MediaStream "active state". [1] http://w3c.github.io/mediacapture-main/archives/20150925/getusermedia.html
The proposed patch for https://webkit.org/b/151346 skips two tests and refers to this bug.
https://webkit.org/b/151346 mentioned above was closed as a duplicate. New bug is: https://webkit.org/b/151360
Spec update is under way https://github.com/w3c/mediacapture-main/pull/291
The tests fast/mediastream/MediaStream-add-ended-tracks.html and fast/mediastream/MediaStream-add-remove-tracks.html are marked as passing on the Mac port. On GTK+ fast/mediastream/MediaStream-add-ended-tracks.html timeouts and fast/mediastream/MediaStream-add-remove-tracks.html fails with this diff: --- /home/clopez/webkit/webkit/layout-test-results/fast/mediastream/MediaStream-add-remove-tracks-expected.txt +++ /home/clopez/webkit/webkit/layout-test-results/fast/mediastream/MediaStream-add-remove-tracks-actual.txt @@ -49,19 +49,10 @@ PASS stream2.getVideoTracks().length is 0 *** active attribute is still true (until event is fired) -PASS stream2.active is true - -Stream2 is inactive. -PASS stream2.active is false - -*** add non-ended track -PASS audioTrack.readyState is not "ended" - -*** active attribute is still false (until event is fired) -PASS stream2.active is false +FAIL stream2.active should be true. Was false. Stream2 is active. -PASS stream2.active is true +FAIL stream2.active should be true. Was false. PASS successfullyParsed is true TEST COMPLETE Updated the GTK+ expectations file in r210841: <http://trac.webkit.org/changeset/210841>
Created attachment 435601 [details] Patch