GGI backend success

Whats happened to this list ?

Anyway I have progressed further with the GGI backend - testvidinfo now
reporting the current mode as 8bpp. an update function, setcolors, listmodes
and setmode
are all working to a degree. Unfortunately getting a pointer to the video
memory is
difficult as ggi sometimes requires that you must aquire this before
accessing the memory
directly (every time you want to do it, then release it afterwards). This is
how I propose to get this working as of now ; (1) if no aquire is required,
provide the directbuffer, (2) if aquire is required
provide a shadow buffer, then blit to the ggi buffer in updaterects(). The
second situation is not good as the buffer may be in an unusual format
(modeX), or may already be double buffered (can’t think why now?).

GGI also seems to be able to emulate any mode - all the modes I’ve tried
return available as of now. Is this a problem for SDL ? (I know it can’t be
that fast)

Where should I send my src/linux/ggi directory to, to make it part of SDL?
(As of 0.10.0 the only change in makefiles is in make/linux/rules, adding GGI
to the possible list
at the end of the file, I believe)

Joel Utting wrote:> hmm, it seems to be working now (gave me a user unknown message back on a

previous message I sent) so I’ll try to repeat what I said in my first

I have been working on the GGI backend for SDL, and after much messing
about, I now have the libSDLggi loading and initing GGI, as well as
setting a default visual. It should be straightforward to implement the
rest of the basic functions in sysvideo, but I haven’t even looked at
sysevents yet - does this need to be done as well (using libgii events) ?

Is the svga backend still being working on ? It’s easier to understand
then the x11 one.

Bye - Joel.