Re: Ogle on Solaris 8, Major success and a little tiny problem..

Hello Hakan,

> We don't have such a cool graphics card here (or machine for that
> matter) (We 'just' SubBlade 1000 1*600MHz with Creator3D or Elite3D
> m6).

First of all, thanks for your reply. Well, the Creator3d and Elite3d should
also exhibit this problem if you manage to use the default configuration.
One way to try this is to escape to a command line login from the CDE
dtlogin program and start openwindows with the default 8bpp root window:

/usr/openwin/bin/openwin -dev /dev/fbs/ffb0 defdepth 8

Verify with xdpyinfo that you still get 1,8,24 depths with 8bpp being the
default. That should reproduce the problem.

Also, about the patch.. perhaps we should protect these lines with
"#ifdef __sun___".. What do you think? I do think that it is portable since
it's taken from xmms and is just plain gtk.

> I've gatherd that the XVR-1000 and above line of graphics cards might
> give better preformance if one doesn't use mediaLib to scale the image
> and XShmPutImage it but instead use a OpenGL texture maped output.
> mplayer has such anoutput mode and hopefully this could be adapted to
> Ogle without to much work.

That's interesting.. I didn't know that OpenGL pixel rendering could be
faster than XShm on that machine... At least, can we keep both versions of
the routines together so we get the choice from ./configure when/if you
introduce OpenGL pixel rendering?

I have another question: Since I am compiling with gcc-2.95.2 from the Sun
Software Companion, my ogle libs get linked to the sparcv8plus+vis2 version
of the Medialib.. Would I gain something by using a 64bit compiler and the
sparcv9+vis2 version of Medialib? If you any idea...

> --
> Håkan Hjort

Vincent S. Cojot