On Sat, Mar 16, 2002 at 05:46:04AM +0100, [EMAIL PROTECTED] wrote:
> This closes the letter of bug 341, but IMO not the spirit -- how should
> we handle a dead soundserver?  Die gracefully immediately?  On startup,
> I think this should be the case.  Anyway, that's easy to fix later.

Argh, make that bug #340.   Silly typo.

But regarding bug 341 -- It supopsedly crashes when there are too many
active voices -- which doesn't make much sense to me since this code is
single-threaded (and shouldn't trigger race conditions).  I'm
really not sure why it would crash.   And ItWorksForMe(tm). 

There are two causes for the driver to run out of voices:

1) Buggy songs that don't send proper note-offs (eg HQ.99)
   - Every "out of voices" message directly follows an improper note-off.
   - It may still be a driver bug.  I'm pouring over the MIDI data.  :)
2) We don't have a true voice allocation/priority scheme in place.
   - Which would sorta solve (1) anyway.

I may try to further mangle the AdlibEmu driver to get a full 16 voices,
which may render both problems moot in some cases.  But who knows.

I'd prefer to not bother with true voice allocation, because it only
affects the adlib -- real midi devices do this for us, and if you're
using a GM softsynth, they have obscene numbers of channels. 
So if we never program a real adlib, I say screw it.  :)

It also turns out that the cc 0x7b (allstop) isn't handled, but it's not
being called anyway.  I just implemented it for the adlibemu driver;
all others are true midi and handle it implicitly.

BTW, bug #302 should have been fixed long ago -- can we verify this?

 - Pizza
-- 
Solomon Peachy                                    pizzaATfucktheusers.org
I ain't broke, but I'm badly bent.                           ICQ# 1318344
Patience comes to those who wait.
    ...It's not "Beanbag Love", it's a "Transanimate Relationship"...

-- Attached file included as plaintext by Listar --

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE8k17iysXuytMhc5ERAhy9AJ0fjHtOIfu0Xlzskecv4PqLUJe0+wCdExWC
ZVKVjx1soRsdQHSakiH1zbs=
=O+AN
-----END PGP SIGNATURE-----



Reply via email to