Vorbis dev - Jun 2001 - ogg123, buffering, tag

I was just about to commit a big buffer overhaul when I realized that
it would probably break a lot of stuff and I'd get yelled at, so it's
in the kcarnold_work tag (or will be momentarily...). Have a look at
it, try to make it deadlock. If I messed up the CVS tagging, also let
me know -- I know only enough about CVS to be dangerous :)

This should also reduce CPU usage noticably. The reader and writer
process both spend a lot less time doing nothing now.

Verbose progress is not as responsive now (it should actually be more
responsive on a slower computer...) as a side effect (when buffering),
but since it's inaccurate anyway (and will probably stay that way),
nobody should really care :)


-- 
Kenneth Arnold <ken@arnoldnet.net> / kcarnold / Linux user #180115
http://arnoldnet.net/~kcarnold/



<HR NOSHADE>
<UL>
<LI>application/pgp-signature attachment: stored
</UL>
-------------- 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/20010622/4f256eb4/part-0001.obj

Let's put things a bit diffferently ...

kcarnold_work contains my latest fixes, hacking, etc.

If you would PLEASE test it on your system, especially with the
buffering mods, and try to nail down the seemingly last case of buffer
deadlock (on buffer_flush -- just hit SIGINT with buffer enabled and
see if it freezesq)? My system does not respond well (decode
timing is the key) when I enable the DEBUG_BUFFER flag -- the deadlock
vanishes because it seems to be all about timing. If you can figure
out what's happening in that deadlock on your system, please fix it if
it's trivial, or mail me back and I may be able to do something about
it.

I'm going away after tomorrow for 3 weeks, and during that time I
won't have much time / connectivity to hack on stuff. If there are any
significant things you want in ogg123 before I go, speak up now.

btw, kcarnold_work branch has, among other stuff:
 * buffer optimizations (but leading to that deadlock)
 * play every nth frame and play every frame n times, just like
   mpg123's d and h options (IIRC).
 * by the time you read this probably some more random hacking

thx,


-- 
Kenneth Arnold <ken@arnoldnet.net> / kcarnold / Linux user #180115
http://arnoldnet.net/~kcarnold/



<HR NOSHADE>
<UL>
<LI>application/pgp-signature attachment: stored
</UL>
-------------- 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/20010622/eaaf16b7/part-0001.obj