I've always suspected this code was racey, and bug 1133064 confirms it. Once we land bug 1135170 and bug 1135785, we'll be in a better place to tackle this.

The previous setup is wacky, and can cause the notification to reach the state machine before the promise rejection, which causes us to stall intermittently. We also take the opportunity to be a bit less trigger happy when we fire it in MediaSourceReader.cpp.

This collided pretty hard with edwin's work in bug 1134434, so I had to rebase it and run it through try again: https://treeherder.mozilla.org/#/jobs?repo=try&revision=fea92dd2b6fc https://treeherder.mozilla.org/#/jobs?repo=try&revision=b6b51350e864 I've filed bug 1146485 to get this stuff fixed up nicer.

remote: https://hg.mozilla.org/integration/mozilla-inbound/rev/05f5e4ea3d54 remote: https://hg.mozilla.org/integration/mozilla-inbound/rev/eb2755790b78

https://hg.mozilla.org/mozilla-central/rev/05f5e4ea3d54 https://hg.mozilla.org/mozilla-central/rev/eb2755790b78