On Thu, 21 Sep 2000, Andreas Beck wrote:
> > > LibXMI
> > I have only had a brief look at XMI. From what I have seen you have done a
> > beautiful job with it and I will certainly take a closer look at it.
>
> Please do so. Maybe this is where your function fits in.
>
> BTW: Jon - I haven't looked at LibXMI for long. Is it a full extension
> (rendering libs and all ?)
I haven't defined any special intrafaces yet, just the set of
exposed API functions.
> If so, you might want to consider adding
> a few utilizations of ACCEL_DRAWTRIANGLE to the kgi-command renderer of
> it.
There are no accelerated targets yet. Well, there is a skeleton X
target, but it doesn't work properly yet. A kgi-command renderer would
need a whole set of sub-targets which would be able to handle the
different chipsets, with the primary target only handling the basics like
resource management, framebuffer mapping, and the kernel<->user
communications pipe (/proc vs ioctls, etc).
> Might make it rock on Permedia chips :-).
Glide would probably be the easiest XMI target to write for.
Clear, clean, simple yet powerful, and has most common 3D accelerated
functionality. A proper KGI driver will make the Glide driver look like a
toy, but will be non-trivial to do properly since even KGI/Dali is pretty
complex. KGI/Degas takes that complexity to a whole new level, and thus
will be even more difficult to do properly. I don't think we should start
out there, but with something a bit easier |->.
Jon
---
'Cloning and the reprogramming of DNA is the first serious step in
becoming one with God.'
- Scientist G. Richard Seed
