The point would be: SDL can be retargeted at a later date, to use GGI, or
SVGALib, or something else, and now you’d have an X server that would run on
that back end. And yes, it would be good to have a free, featureful X server
for Win32 (the free ones available aren’t open source, and don’t support all
necessary features).
I suggest starting with the X server built into VNC
(http://www.uk.research.att.com/vnc/), which currently writes to an invisible
frame buffer, which is then transmitted to connected clients upon request.
However, I think it could easily be retargeted to SDL.
Or, you could start with Xvfb (X Virtual Frame Buffer), which also targets a
memory-based virtual frame buffer. This should be easy to retarget to SDL.
Then, if we got SDL working on SVGALib, we might have a lighter-weight X server
for x86-based UNIX. Also, it would then be a matter of recompiling to bring up
a new, free, X server for Win32. Which would allow more people running that OS
to discover the power of X, and run various X apps locally, with decent
performance, for free.
Warren
______________________________ Reply Separator _________________________________Subject: Re: [SDL] XServer for SDL ?
Author: at internet-mail
Date: 8/26/99 12:28 PM
what do you think of an XServer which uses SDL instead of own drivers ?
i thinking about hacking in the X code
(e.g. implement a faster communication channel than sockets
(ipc,shmem,…) )
Starting with XFree86? Talk about code bloat…
The new Xfree86 4.0 is supposed to separate out low level drivers from
the x protocol stuff so it will be easier to add support (accelerated in
2d and 3d) for new hardware as it comes out. Working within that
context I could see it being feasable doing an SDL driver. But what would
it work under? X? So you’d have an X server running on an X server? Or an
X server running on win32? I guess I’m wondering what would be the point.
-Dave