From: Hartmut Niemann <niemann@cip.e-technik.uni-erlangen.de>
  To  : ggi-develop@eskimo.com
  Date: Tue, 21 Jul 1998 08:34:33 +0200 (MESZ)

Re: new db/graphtype

> 
> On Mon, 20 Jul 1998, Andrew Apted wrote:
> > An 8-bit index into a table of ggi_color.  That's the abstraction that
> > LibGGI provides via ggiSetPalette() & ggiGetPalette().  Who cares what
> > the "real" palette is ?
> 
> Anyone that has to do complex shading/dithering, as they won't be able 
> to just trust ggiSetPallette(); way too primitive.  If such
> routines run under the illusion that the pallette is really
> as precise as ggi_color, that will result in funny moire type effects
> which could be avoided.  Most of this kind of code will be internal to 
> libGGI 2D and 3D, but some applications will need to roll their own 
> customized functions even when they have 2D and 3D handy.
One way that should work is setting the palette and then reading it back,
IF we define 'sane behaviour' to be that GetPalette does not return the cache
but the actual CLUT settings or that the cache is 'resolution correct', i.e.
if you have a 6-6-6 bit clut, all values are &0xfc00'ed.

Then you can set the palette , and read what you got -- you can't do anything
about the hardware anyway in the program...

IMHO a direct way of manipulating the palette is overkill. A correct cache
is enough.
> Brian S. Julin
Hartmut

--  
Hartmut Niemann   --   niemann(a)cip.e-technik.uni-erlangen.de
http://cip2.e-technik.uni-erlangen.de:8080/hyplan/niemann/index_en.html [/ggi]