linuxaudio.org

On Tue, May 24, 2011 at 06:16:49PM +0200, Roberto wrote:

quoted text
> On Tue, May 24, 2011 at 04:40:39PM +0200, Jörn Nettingsmeier wrote:

> > On 05/24/2011 01:12 AM, Eric Steinberg wrote:

> > >Thanks, Paul.  If jack had not crashed, shouldn't rotter have continued

> > >to record?  Is there any way to establish why recording was interrupted?

> > >  This happened about a week after I had started recording on this system.

> >

> > interesting usecase. i never run jack for that long... but i guess

> > some people who are using it for museum exhibits do.

>

> It could be anything actually, but the first thing that went to my mind

> was a 32bit integer overflow (sample count maybe) somewhere in the audio

> software chain...

I thought about that too, but then I started calculating:
4,294,967,296 (unsigned int of 32 bits) / 48000 (seconds)

/ 360 (hours) / 24 (days) = 1.035630617
So if it would be sample count, (jack_nframes_t),

it would overflow after approximately one day, I guess.
greetings,
lieven

_______________________________________________

Linux-audio-user mailing list

Linux-audio-user@lists.linuxaudio.org

http://lists.linuxaudio.org/listinfo/linux-audio-user