[vlc-devel] segfault in mpgatofixed32.c

[jpd at videolan.org]

On Sun, Jul 12, 2009 at 05:53:35PM +0200, Laurent Aimar wrote:
> > Looks like p_out_buf->i_nb_bytes is a bit big compared to p_out_buf->i_size.
>  If possible, a check using valgrind would be a plus.

I can confirm an eerily similar crash when transcoding-then-duplicating,
depending on input format. The difference between valgrind and no valgrind
was the amount of audio played before it'd hit, not the place of the crash.

I'll see if I can reconstruct the sout chain that got me there.