STR: Generate call ID, paste it into window, hit call, share camera/mic, accept call, end call from small window. (linux debug) The left-over audio is fragmentary; bits of the ringtone sometimes, just occasional pops at other times. It continues, but eventually stops. It appears it will stop when I navigate away, and then hit CC and then GC in about:memory. 100% repeatable on my inbound debug build from this morning (and yesterday inbound as well). Not sure who would be appropriate so cc'd a bunch

I've heard the problem on Android and Gonk as well.

Actually, it's a combination of the ring tone and the last thing the party at the other end said aid aid id d d d d.

I see this in gum_test.html, so it's a core issue - likely a regression from bug 848954 stopMedia() calls mozSrcObject.stop() (which seems to work), and then does mozSrcObject = null, which seems to not fully work (anymore). padenot is on PTO, so we'll have to dig this one out without him.

Fairly certain this is related to bug 848954, which means 34 & 35 only

I believe I am running into this on my Android device (Nightly 10/02), e.g, on stopMedia for the audio test on http://mozilla.github.io/webrtc-landing/gum_test.html, I get a never-ending loop of audible audio (sounds like an audio hang on an old frozen PC machine)

Comment on attachment 8499219 [details] [diff] [review] Stop audio sources from continuing to play garbage after being stopped The cause was that if all the audio streams were finished (but not removed yet) they would generate 0 ticks out output - and the Mixer callback would never be called, leaving us to feed the buffer back to the system with whatever was left over in the scratch buffer and a bunch of garbage (which might be previous audio, or distorted previous audio). This clears any data we didn't actually write to.

Comment on attachment 8499219 [details] [diff] [review] Stop audio sources from continuing to play garbage after being stopped Whomever can get to it. Need it for uplift to Aurora/34. Thanks

(In reply to Randell Jesup [:jesup] from comment #8) > Whomever can get to it. Need it for uplift to Aurora/34. Thanks I suppose you forgot about this?

[Tracking Requested - why for this release]: regression

Comment on attachment 8499219 [details] [diff] [review] Stop audio sources from continuing to play garbage after being stopped Approval Request Comment [Feature/regressing bug #]: bug 848954 [User impact if declined]: Random audio for a random time after ending a getUserMedia stream [Describe test coverage new/current, TBPL]: Well tested on 35 and 36 [Risks and why]: very low risk; zero out buffers when data is requested if needed. [String/UUID change made/needed]: none

Comment on attachment 8499219 [details] [diff] [review] Stop audio sources from continuing to play garbage after being stopped Beta+

[Blocking Requested - why for this release]: Can cause audio to continue to play random data (usually somewhat trashed previous-audio-data) after ending a call or audio stream. Simple, safe fix.

Comment on attachment 8499219 [details] [diff] [review] Stop audio sources from continuing to play garbage after being stopped NOTE: Please see https://wiki.mozilla.org/Release_Management/B2G_Landing to better understand the B2G approval process and landings. [Approval Request Comment] Bug caused by (feature/regressing bug #): bug 848954 (landed in 34) User impact if declined: Can cause audio to continue to play random data (usually somewhat trashed previous-audio-data) after ending a call or audio stream. Testing completed: landed many weeks ago on desktop 35 Risk to taking this patch (and alternatives if risky): Simple, safe fix. String or UUID changes made by this patch: none