#10183 NORM Not Tri: Record-82 crashes on os204 while saving a just-recorded audio clip

[Zarro Boogs per Child]

#10183: Record-82 crashes on os204 while saving a just-recorded audio clip
---------------------------------------+------------------------------------
           Reporter:  mikus            |       Owner:  dsd                              
               Type:  defect           |      Status:  new                              
           Priority:  normal           |   Milestone:  Not Triaged                      
          Component:  record-activity  |     Version:  Development build as of this date
         Resolution:                   |    Keywords:                                   
        Next_action:  diagnose         |    Verified:  0                                
Deployment_affected:                   |   Blockedby:                                   
           Blocking:                   |  
---------------------------------------+------------------------------------

Comment(by Quozl):

 Replying to [comment:9 mikus]:
 > Replying to [comment:8 Quozl]:
 > > (sugar-launch org.laptop.RecordActivity)
 > What that does is put output to Terminal (which acts as a console),
 rather than to .sugar/default/logs

 Yes.  What it does that the logs don't, is show the ''Segmentation fault''
 message.

 > > (strace)
 > Had to make several attempts.  On the XO-1.5 (os204), with python-psyco
 present, "plain" launching of Record-82 consistently resulted in the
 #10183 problem (Terminal showed segment fault).  But with 'strace', I was
 NOT able to reproduce the #10183 problem on the XO-1.5.

 Yes, I also encountered that ... running with strace reduces the frequency
 of the symptom.  Creating movement in the field of view of the camera
 increases the frequency of the symptom.  This suggests to me that the
 problem is a [http://en.wikipedia.org/wiki/Race_condition race condition].

 > Am attaching the tail end of the console output, plus the whole 'strace'
 recording.

 Thanks.  Given that the symptom is easily reproduced, and a gdb backtrace
 points at a GStreamer involvement, I think we probably have enough
 information for a developer to work the problem further.

-- 
Ticket URL: <http://dev.laptop.org/ticket/10183#comment:10>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system