Observed while trying to reproduce bug 870564. STR: 1. open the YT link in bug 870564 comment 0 2. click on the "Play" icon to open the video in the Video app Observed behaviour: sometimes (more specific STR aren't clear, but it seems to be related to either network congestion or CPU usage on the DUT--this happens about 1/3rd of the time) the video frame stops while the audio continues to play. Often when this happens, the video's audio will continue to play to the end, but the video never updates. Expected behaviour: if the audio can play, the video should play as well. Observed on Inari with: - gecko: b2g18:9cac5a5c4efb - gaia: c51082d92f18c4c7cdbc4f96c977be735500562a

This sounds exactly like the behavior we hit when testing the youtube certified app prototype with their HTML 5 mobile site.

As I pointed out in bug 877461 I have the feeling this is becoming quite complex to be solved in 1.0.1 timeframe.

It seems that only video track playback ended by some error.

(In reply to Jason Smith [:jsmith] from comment #1) > This sounds exactly like the behavior we hit when testing the youtube > certified app prototype with their HTML 5 mobile site. So are you able to repro with any regularity? Mike suggests 1/3

(In reply to Alex Keybl [:akeybl] from comment #4) > (In reply to Jason Smith [:jsmith] from comment #1) > > This sounds exactly like the behavior we hit when testing the youtube > > certified app prototype with their HTML 5 mobile site. > > So are you able to repro with any regularity? Mike suggests 1/3 Not consistently, but it's possible. cpearce, myself, jeni, and Sreenidhi have been able to reproduce this with the YouTube certified app. Haven't tried the Gaia:Video equivalent in this bug, but it sounds like the exact same behavior we're seeing in bug 877461. Play a few videos on YouTube and you'll definitely eventually hit this. I'm guessing this bug and bug 877461 are one of the same, but I need someone to confirm. Sotaro - Is bug 876843 and bug 877461 the same issue, even though this bug was tested in Video app running YouTube videos, while bug 877461 was tested with the YouTube certified app?

(In reply to Jason Smith [:jsmith] from comment #5) > Sotaro - Is bug 876843 and bug 877461 the same issue, even though this bug > was tested in Video app running YouTube videos, while bug 877461 was tested > with the YouTube certified app? I think it happens by same reason. It happens often when MediaStreamSource::readAt() is blocked longer time. http://mxr.mozilla.org/mozilla-b2g18/source/content/media/omx/OmxDecoder.cpp#95

Okay, so this is a confirmed dupe then. Most of the discussion on this is happening over in bug 877461, so I'm duping ahead.