From: Nicolai Haehnle <Locinga@pfaffenwinkel.de>
  To  : ggi-develop@eskimo.com
  Date: Sun, 07 Mar 1999 21:26:48 +0100

Re: Sprites in video memory

becka@rz.uni-duesseldorf.de wrote:

> O.K. - I have kept out of that discussion up to now, but ...
>
> > > > 2. No hardware support for Blitting and other features.
> > > It's true, but when you draw sprites you haven't hardware acceleration
> > > at all. I don't know any card that supports blitting with masks... Maybe
> > > my own card supports it and I don't know this :)
> > I'd think most modern cards support this.
>
> Yes. S3 964 were able to do it, so it would be a shame, if newer ones don't.

The riva supports "Hardware blitting with source colorkey" (at least, DX
diagnostic tools say so)

>
>
> > > > and some other thing.  Then you should be able to draw something on VRAM
> > > > using hardware acceleration.
> > You are basicly describing the already existing sub-target here.
>
> Exactly. I really want to know, what you are all argueing about.
>
> You have memvisuals, and you have the subtarget, so where is the problem ?
> Except for the little problem, that you can't specifically allocate just
> one "offscreen area" if you have multiple frames.

The sub-target is a nice thing, but as Brian said the visual structure is bulky
and, which is more important, the application programmer must decide where to
locate the sub-target.The best solution for all this would be to make the
appropriate targets aware of all this, but this requires heavy changes in the
API.
What I will do for now is to write a library that allocates drawmaps for the
programmer, as I've already described in an earlier posting. It won't be a libGGI
extension because it acts as a container for several visuals.

cu,
 Nicolai