Re: G400 Multitexture was (Re: [Dri-devel] demos/fire fogging?)

[Jeff H. <jha...@va...>]

Jeff Hartmann wrote:
> 
> Ville Syrjälä wrote:
> >
> > On Wed, 1 Nov 2000, Keith Whitwell wrote:
> >
> > > Gareth Hughes wrote:
> > > >
> > > > The fog values for at least the ground plane are incorrect for both the
> > > > mga and r128 driver.  Thus, it looks like a core Mesa bug (in the fog
> > > > coord stage perhaps?).
> > > >
> > > > Keith, do you want to verify this?  I wouldn't mind looking into fixing
> > > > the problem.  I will submit a bug report now.
> > >
> > > It's not a bug, it's simply an artifact of per-vertex fogging vs. per-pixel
> > > fogging.  This is one good reason for trying to get fog tables working on the
> > > r118.
> >
> > What about G400 table fog? It would be nice to use the second texture unit
> > for something.
> 
> This does not work when the second texture unit is in use.  We could
> special case fog so it works w/ the table when only the 1st texture unit
> is being used, but we will have to use per-vertex fogging in the
> multitexture case.
> 
> >
> > BTW is the multitexturing on the G400 disabled by simply disabling the
> > extension? Something like "if (0 & ...)" in some mga_XXX.c. FYI I removed
> > the "0 &" and the extension gets reported and quake3 runs fine.
> 
> I'm actually going to commit a fix for the multitexture bug quite
> shortly.  You are right that the extension is just disabled, but very
> soon it will not be ;)
> 
> About quake running fine:
>         The multitexture bug is very weird.  On most machines I've been testing
> on it does not happen in full screen mode.  It also happens much more
> often with high resolution modes in quake3.  It can be VERY visable, or
> hard to reproduce depending on what modes you are running quake in, how
> fast your machine is, etc.
> 
> -Jeff

Okay I checked in the fix for the multitexture bug.  Multitexture is
stable again (woohoo!)

-Jeff