Re: [xine-devel] v4l patch2, take 2 ;)

[Jeroen A. <j.a...@it...>]

Miguel Freitas wrote:

>On Mon, 2003-09-15 at 04:01, Jeroen Asselman wrote:
>  
>
>>>what is this patch for? to handle the no input signal case?
>>>      
>>>
>>Yes, if there is no input signal my grabber card doesn't always deliver 
>>frames, in which case xine would now and then freeze (this still 
>>happened with rc0), and xine needed to be killed. This patch should send 
>>an end of stream when there is no input signal and thus avoid freezing.
>>    
>>
>
>is it still needed? (now that "undocumented feature" was removed ;-)
>  
>
Apparently it is. Xine rc1 can still freeze when  there is no valid 
input signal. This doesn't always happen. It doesn't seem to be the 
input plugin that freezes xine, but the xine-engine itself seems to 
freeze somewhere. I am unable to locate where it happens though. So you 
could say the patch is a workaround. Is there a key combination I can 
press to force xine to generate a core dump? Than I should be able to 
see where it stalls. I can't just attach with gdb because there is no 
gdb on the machines where the TV card is inserted and tested with. The 
development is done remotely.

The attached patch fixes a segfault when trying to listen to the radio 
though. Please apply to CVS. It is only a one line change :).

- Jeroen