Date:      Thu, 10 Feb 1994 12:40:05 MET
From: "X Appeal Info" <xappeal AT xtreme DOT it>
Organization: Xtreme s.a.s. - Livorno, Italy
Reply-To: "X Appeal Info" <xappeal AT xtreme DOT it>
To: ""John E. Davis"" <davis AT amy DOT tch DOT harvard DOT edu>
Cc: djgpp AT sun DOT soe DOT clarkson DOT edu
Subject:   Re: ATI Wonder Card

On Wed, 9 Feb 1994 11:11:05 -0500, "John E. Davis" <davis AT amy DOT tch DOT harvard DOT edu>
wrote:

>    I have a compaq 386 with an ATI Wondercard manufactured in 1989.
> Yesterday, I installed a demo version of XAPPEAL, an X11 server for MSDOS.

Which version? The current one is 1.4.2.

> The documentation indicated that it was written using DJGPP.  The only
> problem I had with the software is that the video would not sync well.  I am
> using an NEC Multisync 3D monitor so I do not believe that this is due to
> the monitor.  Rather, I believe that the problem lies with the graphics
> driver. The drivers that come with the program includes one called
> ativga.gfx which support ATI 18??? chipsets.  Is the ATI Wondercard of this
> type?  If not, does anyone have a driver for it?

Our drivers have the extension .GRD up to version 1.3, and .GXN in 1.4.0
and later versions. Both are based on the DJGPP drivers (old and new
style), but with extensions to allow setting a virtual resolution
different from the real one and doing panning. We never used the .GFX
extension.

>    I reviewed the documentation of the graphics drivers that come with DJGPP
> and it appears that DJGPP can only support graphics cards that are
> compatable with its paging mechanism.  Is the ATI Wondercard such a card?  

Yes, and it should be handled correctly by our driver. But maybe
there is some mistake, we did not test it on all types of ATI chips.

Then, on Wed, 9 Feb 1994 21:41:25 EST, "Bob Babcock" <babcock AT cfa DOT harvard DOT edu>
wrote:

> The ATI VGA Wonder used the 18800 chipset.  But, based on my experience,
> there are different ATI cards, possibly using the same chipset, which require
> different code for bank switching.  The symptom of using the wrong bank
> switching code is that, as the screen is being updated, you'll get a horrible
> out-of-sync glitch in screen modes larger than about 640x480(x16 color?).
> It's often bad enough to shut down my multi-sync monitor.  I've just received
> the 32-bit version of GLE (a graphics package) which uses djgpp, so I may
> soon face the same problem.  If so, I'll try to create a fixed driver file.
> But there are more important things on the queue first.  If you'd like, I can
> send you an example of working bank-switching code from QPEG (a JPG display
> program).

And on Wed, 10 Feb 1994 00:18:18 -0500, "John E. Davis" <davis AT amy DOT tch DOT harvard DOT edu>
wrote:

>     Thank you very much.  I just received the following assembly language
> routine from jmiller AT umdsun2 DOT umd DOT umich DOT edu.  Unfortunately, I do not really
> know quite what to do with it.  Do I simply assemble it and rename it to
> something like `wonder.gfx'?

No, since the entry points are somewhat different. We hope to be able
to standardize on a new driver format without proprietary extensions
and with all the features we need, but this has not been possible
before the current release. The .GXN format is similar to the .GRN
one, but there are some additions.

We are currently looking at the code of the driver you posted,
together with all the information we have from different sources
about the ATI chips; I suggest a follow-up by email between us
without bothering the DJGPP mailing list readers. Please get in
contact with us at this address.

-- 
Giovanni Novelli
Xtreme s.a.s. - Livorno, Italy
xappeal AT xtreme DOT it