Re: [MiNT] New MiNT distribution is n.. ..pro

[Johan Klockars <rand@cd.chalmers.se>]

> > Really ? Well, if so, fVDI really seems to be a good solution, for the VDI

I made a new beta version (v0.95beta) available a couple of days ago on
my ftp site, in case you want to try it out. This includes drivers for the
Eclipse/RageII as well as the normal Atari monochrome and 16 colour modes.

It's been six months since the last fVDI engine version that the bitplane
drivers were compatible with, but they seem to be mostly fine again now (they
still have a few problems with Bezier curves and areas, but I expect to fix
that up in a couple of days).
Any bug reports or comments are very welcome!

> Mesa has already been ported, but I don't think it supports any
> OpenGL-hardware. If it doesn't then it doesn't make much sense in using it.

At least nVIDIA has Linux code for hardware accelerated OpenGL available,
and I think that might be true for 3Dfx too. Since I bought an nVIDIA RIVA128
PCI card some time ago (I'm looking for a cheap 3Dfx 2D/3D card too), I'm
definitely going to try to do something about that sooner or later.

S3 apparently also makes enough information about their (at least older) 3D
hardware available, and possibly Matrox. So far I haven't seen anything
useful for the ATI Rage, though.  :-(

> >         It would be great if someone could make a listing of things to
> > do if MiNT was to be ported to the RioRed platform...
> 
> Write the boot-strap code, write the 68k-emulator, possibly port MiNT
> to PPC (or run it on the emulator), write drivers for fVDI (unless you stick
> with ATI Rage), write hardware-drivers for MiNT... More than enough ;-)

The boot-strap code should be mostly the same for any OS and there are
already several 68k-emulators available that could be used.

A primitive fVDI driver (not counting hardware initialization) can be thrown
together in an afternoon, but that will of course be quite slow. I turned off
all acceleration in the Eclipse/RageII driver yesterday for some tests, and
while it was _very_ slow, it was actually somewhat usable (the main problem
being that the mouse drawing was so slow that the interrupts acted up).

Writing drivers for various hardware would be the most time consuming task,
and that's the main reason why I personally don't believe in actually
porting TOS/MiNT to hardware that already has an OS (especially if you'd end
up running an m68k-emulator, anyway).
While I understand Jo Even's dislike for running things on top of another
OS, it makes things _much_ easier and in some cases it's even the only
possible way to use certain hardware (it's not too uncommon that there simply
is no low-level hardware information available).

-- 
  Chalmers University   | Why are these |  e-mail:   rand@cd.chalmers.se
     of Technology      |  .signatures  |            johan@rand.thn.htu.se
                        | so hard to do |  WWW/ftp:  rand.thn.htu.se
   Gothenburg, Sweden   |     well?     |            (MGIFv5, QLem, BAD MOOD)