It's possible for ended to still be true in the seeked event handler after seeking to 0 in the ended event handler. In this case, mEnded in nsHTMLMediaElement can only be reset to false in ResourceLoaded, which may not run soon enough. According to the spec, ended should be false if currentTime is not the end of the media. The seek algorithm changes currentTime during the synchronous part of the algorithm, so the following should be true (but may not be with the current code): function ended() { is(v.ended == true); v.currentTime = 0; is(v.ended == false); }

Similar problem with v.seeking, according to my reading of the spec this should be true immediately after the setting of currentTime returns. The current implementation only returns true once the decoder has transitioned to the seeking state, which may happen some time later.

Drop the mEnded flag in the media element, and instead ask the decoder if playback has ended. Change IsSeeking() to report if a seek has been requested by also checking the next state of the decoder. Also silence a warning in the Wave decoder where an event may be null when seeking after playback end. Unfortunately the Wave test for this bug depends on yet another bugfix which is at the other end of my patch queue. I'm in the process of untangling that from a big ball of fixes and will file a new bug with patch for that and link up the dependencies appropriately.

Same patch, but removes the test because it depends on another bug being fixed. I have a bunch of patches like this, where the testcase tests the patch but breaks because of some other bug. I'm going to open a new bug and attached a patch to that included all of the testcases, and that can land on top of the stack of bugfixes. SR carried forward from roc.

Comment on attachment 352674 [details] [diff] [review] patch v2 + // Call on the mai thread only. 'main' typo. r+ with that fixed.

Make patch less kawaii. R+SR carried forward.

This needs to be applied on top of bug 469255.

Rebased.

Comment on attachment 352876 [details] [diff] [review] patch v2.2 [Checkin: Comment 8 & 9] http://hg.mozilla.org/mozilla-central/rev/3d8e98242012

Comment on attachment 352876 [details] [diff] [review] patch v2.2 [Checkin: Comment 8 & 9] http://hg.mozilla.org/releases/mozilla-1.9.1/rev/1fcfd425228f