Re: [AD] Allegro 4.1.14 WIP

[Chris <kcat@xxxxxxxxxx>]

Eric Botcazou wrote:
> > Umm.. if I may ask, how come this release was so abrupt?
>
>
> We briefly talked about the rationale a couple of weeks ago.  Basically, it 
> is aimed at being somehow a reference release for MacOS X.

Ok, I can see that. It just seemed a little abrupt to me since there 
really wasn't much recent talk about it. I just wanna make sure we don't 
jump the gun on a 4.2 release before it's properly discussed, or something.

> I saw the mixer patches.  But I'm not sure I can make a judgment on them 
> because I know next to nothing about the sound stuff.  And I'm a little 
> worried about ditching the proven mixer at this point.  Would it be feasible 
> to have the two mixers live together and be able to switch at run-time?

Well, the problem is not really with having both mixers. The problem is 
is that the DX mixer works completely bass-ackwards to how the other 
sound drivers work. The way the DX mixer requires things to work is for 
the sound API to call into the sound drivers, which would in turn call 
into Allegro's mixer core as needed. This works, but because of the API 
and core mixer design, it doesn't leave much room for flexibiility. I'll 
write a more detailed plan of attack for this later, but basically what 
I want to do is have the sound API call into Allegro's mixer, and have 
Allegro's mixer call into the sound drivers as needed. This makes sure 
nothing gets bypassed/hacked around, the sound API can be integrated 
closer to the mixer (improving efficiency and reducing code), and makes 
it easier to implement things in software where hardware acceleration 
isn't possible (meaning it'll be easier to support hardware voices while 
maintaining software voices).

- Kitty Cat