[mythtv] Strange backtrace from mythtranscode crash

[Paul Gardiner]

On 01/10/2011 18:36, Jean-Yves Avenard wrote:
> Hi
>
> On 2 October 2011 03:43, Paul Gardiner<lists at glidos.net>  wrote:
>
>> Note also the value of "this" for level #1. Anyone seen anything like
>> this before? I wonder if something previously has made a random write
>> to memory and screwed up the stack maybe. If that were the case, how
>> is it best traced?
>
> In
>              bufsize += len - freebuf;
>              unsigned char *tmpbuf = new unsigned char[bufsize];
>              memcpy(tmpbuf, audiobuffer, audiobuffer_len);
>              delete [] audiobuffer;
>              audiobuffer = tmpbuf;
>
> May be worth tracing bufsize vs audiobuffer_len (though by the look of
> things it should always be greater)

Thanks. It's ab_count going out of range. I'll see if I can fix it
and put up a patch.

Cheers,
	Paul.