[mythtv] MythFrontend crash watching live TV

[J. Donavan Stanley]

Christian Hack wrote:

>>-----Original Message-----
>>From: mythtv-dev-bounces at mythtv.org 
>>[mailto:mythtv-dev-bounces at mythtv.org] On Behalf Of J. Donavan Stanley
>>
>>Christian Hack wrote:
>>
>>    
>>
>>>Got the attached crash while watching live TV. A scheduled 
>>>      
>>>
>>recording was due
>>    
>>
>>>to start at 18:00. The "Do you wish to record .. etc" OSD 
>>>      
>>>
>>came up. I simply
>>    
>>
>>>ignored it and let it count down. When it came to start the 
>>>      
>>>
>>recording you
>>    
>>
>>>can see at the beginning where it's trying to open the file 
>>>      
>>>
>>which it claims
>>    
>>
>>>doesn't exist (yet). Sometimes this works and the number of 
>>>      
>>>
>>retries doesn't
>>    
>>
>>>reach zero. I don't know why there is such a delay. The 
>>>      
>>>
>>system is XP2400
>>    
>>
>>>using DVB cards.
>>>
>>>If the retries gets to zero it ends up in a seeming infinite loop of:
>>>      
>>>
>>If this using current CVS?  If so can you try adjusting the usleep on 
>>line 402 of RingBuffer.cpp up a bit to see if that allows it to 
>>recover?  The problem you're seeing is due to the fact that 
>>when we open 
>>a new RingBuffer we're not able to read any data from the file, and 
>>never will be allowed to read from that file handle.  So we 
>>attempt to 
>>read from the file before returning the handle back up the chain.
>>
>>The recent changes to CVS that start the recorder up ahead of 
>>the player 
>>might also help, though I've not looked to deeply into it.
>>    
>>
>
>Reasonably recent CVS. About a week ago (Oct 19, Australia time). Increasing
>that sleep time won't help it from crashing if the file never appears for
>some reason though. Can we guarantee that the file will appear (eventually)?
>
If  the recording is actually happening yes.  However, the failure to 
open should set an error state in current CVS then we'll need to look at 
that.