From: John Wood <John.Wood_at_royalblue.com>
Date: Fri, 6 Sep 2002 14:04:06 -0400

It you had a stack based at the end of the buffer, then you woudln't get
fragmentation. I'm not suggesting using the buffer as a heap. Most
operations (loading a folder, loading a playlist, doing a search) are
cascading operations and so stack allocation is appropriate.

When you allocate stuff on the stack, it'll shrink the cached MP3 data.
It'll just run out of cache sooner than it did previously.

Unless you're going to tell me that the decoder DMA locks the whole cache
while it's using it and that you don't have control to shrink it while it's
in use? That would put an end to that idea if that's the case :)

-----Original Message-----
From: Linus Nielsen Feltzing [mailto:linus_at_haxx.se]

Have you even thought of how the mp3 buffer handling would be
implemented if the memory was dynamic? Have you considered how it
would handle a fragmented buffer?
Think about it for a while, and imagine the consequences, regarding
the DMA handler, the fast forward/rewind, the time calculations, the
disk reader, the track change detection, and...

/Linus

-- 
Linus Nielsen Feltzing, linus_at_haxx.se on 2002-09-06
*************************************************************************    
This message is intended only for the stated addressee(s) 
and may be confidential.  Access to this email by anyone 
else is unauthorised. Any opinions expressed in this email
do not necessarily reflect the opinions of royalblue. 
Any unauthorised disclosure, use or dissemination, 
either whole or in part is prohibited,.   If you are not the 
intended recipient of this message, please notify 
the sender immediately.
*************************************************************************

Received on 2002-09-06