845928 - crash in libOmxH264Dec when recording a video through the camera app in gaia ui test automation

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: General, defect)

Description

[Jason Smith [:jsmith]]

This bug was filed from the Socorro interface and is 
report bp-109aec06-6c8d-4447-8f73-495c52130227 .
============================================================= 

Our gaia ui test automation is getting failures due to a crash with the URL above when recording a video in the camera app that is five seconds long.

Logcat:

http://qa-selenium.mv.mozilla.com:8080/view/B2G/job/b2g.unagi.gaia.v1-train.ui/130/artifact/logcat.txt

Comment 1

[Jason Smith [:jsmith]]

Mike - Any ideas?

Comment 2

[Mike Habicher [:mikeh] (high bugzilla latency)]

I can't open the logcat link above (FFx says "Server not found")--can you attach the log to this bug?

Comment 3

[Jason Smith [:jsmith]]

Attachment: Logcat

Comment 4

[Mike Habicher [:mikeh] (high bugzilla latency)]

I've looked through that log and it looks like the video recorded properly, and then the camera is closed/shutdown cleanly.  The only thing I see is this:

I/Gecko   ( 1537): ###!!! [Child][RPCChannel] Error: Channel closing: too late to send/recv, messages will be lost

...but by this point the recorder/camera is destroyed.  There's nothing else unexpected in the log--no stack dump, zombie message, SIGanything.

My _guess_ is the crash is in post-parsing of the video, or in a follow-up opening of the gallery.  What are the STR to get to this crash?

Comment 5

[Jason Smith [:jsmith]]

The only STR we've got is from the capture a video test here - https://github.com/mozilla/gaia-ui-tests/blob/master/gaiatest/tests/test_camera.py

Comment 6

[Mike Habicher [:mikeh] (high bugzilla latency)]

Hmm, that script looks okay to me.

Any chance there's a relatively painless way to run this test locally?

Is this 100% reproducible?

Comment 7

[Stephen Donner [:stephend] Not actively reading bugmail]

(In reply to Mike Habicher [:mikeh] from comment #6)
> Hmm, that script looks okay to me.
> 
> Any chance there's a relatively painless way to run this test locally?
> 
> Is this 100% reproducible?

Sorry for the delay; I couldn't get this to reproduce in our automation again yesterday (and we run it often), nor could I, manually or with automation, again, using today's engineering Unagi build.

Comment 8

[Andrew Overholt [:overholt]]

We discussed this in triage today and since it seems like it's not easy to reproduce and is unlikely to be hit by users, we can't even track.  Please re-nom for tracking or maybe even blocking if we can get a better STR.

Comment 9

[BMO Automation]

Firefox OS is not being worked on