Re: Uninitialised storage? (no advance in playlist is a *real* bug)

On Mon, May 26, 2003 at 10:41:12PM +0200, Haavard Kvaalen wrote:
> The behavior could however easily be explained by a problem with the 
> soundcard driver or some kind of transiend hardware error.
> 
> When xmms reaches the end of a song it will wait until the buffer in the 
> soundcard is (almost) empty.  Here it looks like the soundcard driver is 
> still reporting that there is a lot of data left in the buffer, and xmms 
> will never continue to the next song.

I was thinking along those lines. The strange thing was I had stopped and
restarted xmms many times (including recompiles of xmms/main.c with various
fprintf's added) and the behaviour was entirely consistent every time -
until I did the 'make install' as described before. Audio was actually
playing with no problems.

The system is FreeBSD-4.8 compiled with:

device          pcm
device          sbc0 at isa? port 0x220 irq 5 drq 1 flags 0x15

and is running on a Sony Vaio PCG-C1F.

Should it ever happen again, I'll start to probe the return values from the
audio ioctls. For instance, looking at the code, I see that if
SNDCTL_DSP_GETBLKSIZE had returned a silly (negative) value, then the
termination test would always fail:

        if (!oss_used() && (device_buffer_used - (3 * blk_size)) <= 0)
                return FALSE;

> > The other strangeness was when I first installed xmms, the default output
> > device selected was the DiskWriter, and so I was getting no audio and the
> > track time was decrementing at about 3 times real time.
> 
> This may happen if xmms is configured to use a plugin that no longer
> exist.

OK. I recently did a complete O/S re-install, but my home directory was in a
partition which was not wiped. It's possible there was a latent .xmms
directory left over from a long time ago.

Thanks for the feedback...

Regards,

Brian.