[vorbis-dev] Ogg123 buffering

[Aaron Plattner]

Here's a short patch that uses signals to stop and resume the writer thread.  I
don't know if using signals is the best way of doing this, but it certainly
works, and it's short.  There's a race between the writer calling writer_main
and the main thread closing the buffer, but that seems so unlikely as to not
create a problem.  The only problems I can think of are with very short ogg
files or a buffer containing only one chunk, which doesn't work anyway.

Oh yeah... good job to everyone regarding the release.

Aaron Plattner

On Wed, Feb 28, 2001 at 05:38:50PM -0500, Kenneth Arnold wrote:
> No you are not; it's an implementation issue. The present buffer code
> does not have a way to tell how empty or full the buffer is. It is
> still useful for sudden system load or intermittant network traffic,
> but this is something that has been lacking. I'm very caught up in
> lots of other things (as if you all couldn't tell...) so if you can
> fix that, and maybe add a semaphore or other IPC mechanism instead of
> my hacklike spinlock on the location pointers (or even one lock per
> each chunk_t) while you're at it, everyone here can have a big party
> (while I finish my homework). ;)
> 
> Bah, I should maintain my code...
> -- 
> Kenneth Arnold <ken at arnoldnet.net> / kcarnold / Linux user #180115
> http://arnoldnet.net/~kcarnold/


<HR NOSHADE>
<UL>
<LI>text/plain attachment: buffer.diff
</UL>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: buffer.diff
Type: application/octet-stream
Size: 2674 bytes
Desc: not available
Url : http://lists.xiph.org/pipermail/vorbis-dev/attachments/20010228/364cfc24/buffer.obj
-------------- next part --------------
A non-text attachment was scrubbed...
Name: part
Type: application/octet-stream
Size: 233 bytes
Desc: not available
Url : http://lists.xiph.org/pipermail/vorbis-dev/attachments/20010228/364cfc24/part.obj