1251357 - MediaStreamTrack.stop does not stop the track

Status: RESOLVED FIXED

(Core :: WebRTC: Audio/Video, defect, P1)

Description

[msander]

User Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/48.0.2564.116 Safari/537.36

Steps to reproduce:

Open http://mozilla.github.io/webrtc-landing/pc_test.html and start the call.

In the console stop the tracks.
localvideo1.mozSrcObject.getTracks().forEach(function(track){track.stop();})
localvideo2.mozSrcObject.getTracks().forEach(function(track){track.stop();})
 


Actual results:

The video stops playing but the camera is still on.


Expected results:

The camera should be released.

Comment 1

[msander]

Calling mozSrcObject.stop() releases the camera but gives the deprecated warning.

Comment 2

[Mark Banner (:standard8)]

Confirmed as a regression in nightly builds - branches are fine.

Comment 3

[Maire Reavy [:mreavy]]

Thanks for the needinfo.  I've asked jib to do the analysis.

Comment 4

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

Regression range is https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=3eca89d56c0252b12f8c5dadade8e8e76d852258&tochange=218d16a9ddcc3699be2a19bcab7d5f6b7f8e7515

It's probably https://hg.mozilla.org/mozilla-central/rev/1aa33a896ebc

Which is actually a cherry-picked patch from from pehrsons clone work: https://reviewboard.mozilla.org/r/32167/diff/3#index_header

Comment 5

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

I've verified that the commit I mention in comment 4 is indeed the culprit. We can back it out, which will likely bring back the intermittent bug 957691, or I can try to dig in more to figure out what's going on.

FWIW, the workaround for this bug is to wait 20 seconds for garbage collection to kick in, at which point the camera light goes out (I wonder how that plays out in our automated test tree...)

Comment 6

[Randell Jesup [:jesup] (needinfo me)]

I'm sure it's caused by the removal of this: -  void StopTrack(TrackID aTrackID) override
Perhaps that's been replaced elsewhere in pehrsons huge queue; it may be worthwhile to pull that in and see where stopTrack leads to in his full commit.

Comment 7

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

(In reply to Randell Jesup [:jesup] from comment #6)
> I'm sure it's caused by the removal of this: -  void StopTrack(TrackID
> aTrackID) override

Looks like that's it. Good call.

Comment 8

[Jan-Ivar Bruaroey [:jib] (needinfo? me)]

Attachment: MozReview Request: Bug 1251357 - fix regression where last MediaStreamTrack.stop did not turn off the camera light.

Review commit: https://reviewboard.mozilla.org/r/36881/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/36881/

Comment 9

[Randell Jesup [:jesup] (needinfo me)]

Comment on attachment 8724134 [details]
MozReview Request: Bug 1251357 - fix regression where last MediaStreamTrack.stop did not turn off the camera light.

https://reviewboard.mozilla.org/r/36881/#review33509

Comment 10

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/04ee9e8be26f

Comment 11

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/04ee9e8be26f