[ardour-dev] export problems - testers requested

[Jan Depner]

On Sat, 2005-04-30 at 15:56, David Mulcahy wrote:
> On Saturday 30 April 2005 18:54, Paul Davis wrote:
> > Problems with export are one of just two remaining items preventing
> > the release of ardour 0.99. Reports to the mailing list/mantis/#ardour
> > seem inconclusive at present. One user who was having crashes every
> > time he tried to downsample from 48kHz -> 44.1kHz reports that
> > updating his system fixed the problem. Other reports are confusing, or
> > anecdotal.
> >
> > I would really appreciate it if as many as people as possible could
> > try exports on sessions longer than, say, 5 minutes, with and without
> > down and upsampling. These tests should be run with JACK in realtime
> > and non-realtime mode, and you should note which version of JACK you
> > are using. Also report which version of glibc you have, like this:
> >
> 
> 
> Firstly it only happens at samplerate conversion for me 44.1 -> 48 (eg I 
> record at 44.1 and resample to 48 - both 16 bit)
> and at a low period size 128 (2048 is fine).  The following error poped up but 
> not always
> *** glibc detected *** free(): invalid pointer: 0xb3727128 ***
> in fact pretty sure ardour pops up something about glibc in its error window 
> before it shutsdown)
> I will give you more details later but i think it is pretty easily to 
> reproduce once you now how - and it pretty much brings down jack and ardour.
> 
    I just tested this and get a segfault when JACK is running at 128. 
At 2048 , no problems regardless of the settings (up/down sample). 

Running 
/lib/libc.so.6
jack 0.99.61
Ardour/GTK 0.9beta29
   (built using 0.594.0 with libardour 0.890.0)

Jan