[vorbis] Re: Squelch 1.0beta9

[rik at kde.org]

#if Kenneth Arnold
> On Tue, Feb 06, 2001 at 11:10:57PM -0500, Bill Nottingham wrote:
> > Kenneth Arnold (ken at arnoldnet.net) said: 
> > > Then just type 'run'. If you get a weird message about being unable to
> > > read from some memory location (that maps to /lib/ld-linux.so.2 for
> > > me), then you have duplicated the problem I'm trying to track down
> > > with very little success. I've segfaulted gdb several times, but I
> > > have been able to confirm (with strace) that the writer process is
> > > segfaulting, and I don't know why. HELP!
> > 
> > You certianly appear to be calling ao_play without ever calling
> > ao_open on the device, so you end up jumping into the
> > middle of nowhere...
> 
> Ah crud, you're right: somebody (too late; can't remember who)
> committed the open_audio_devices hackish (could be done cleaner)
> reopen-on-bitrate-change, and since the buffer isn't reset after each
> change, or even given notice when the devices are opened in the first
> place, it's dereferencing a pointer-to-nowhere and faulting. The odd
> thing is that I knew this, and in fact was in process of doing another
> batch of cleanups that included a fix, but still forgot that it was an
> issue.
> 
> Okay, problem #1 solved with a "duh, I knew that." That leaves problem
> #2: gdb. I cannot debug ogg123 on my system, but compiling oggenc with
> debug code and gdb'ing that works fine. I haven't conclusively found
> if this is unique to my setup. But explain this rough transcript:
> 
> cd vorbis-tools/ogg123
> cvs -z4 update -dPA
> make clean
> make CFLAGS=-g
> gdb ogg123
> [copyright message removed]
> (gdb) break main
> Breakpoint 1 at 0x80493bd: file ogg123.c, line 100.
> (gdb) run
> Starting program: /usr/src/cvs/xiph.org-cvs/vorbis-tools/ogg123/ogg123
> Cannot access memory at address 0x40017e30
> (gdb) bt
> #0  0x4000d114 in _dl_debug_state () from /lib/ld-linux.so.2
> 
> It's not even getting to main(). gdb bug, ogg123 bug, PEBCAK, or what?
> It does run fine (without buffer) when run normally. This is tested on
> 2.4.0 and 2.4.1, gcc 2.95.3 (may be guilty), gdb 5.0, Debian unstable.
> Any other results? Please confirm that it's not just me.

Works fine for me...

Breakpoint 1, main (argc=1, argv=0xbffff54c) at ogg123.c:100
100         int option_index = 1;

This system is a mix of SuSE 6.4 and some 7.0 packages.

gdb -> 5.0
gcc -> 2.95.2
kernel 2.4.1

I've had gdb do the same to me in the past, but I never figured
out why. Sorry I can't help there.

Cheers,
Rik

--- >8 ----
List archives:  http://www.xiph.org/archives/
Ogg project homepage: http://www.xiph.org/ogg/
To unsubscribe from this list, send a message to 'vorbis-request at xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is needed.
Unsubscribe messages sent to the list will be ignored/filtered.