Re: Scoring visuals in GdkRGB

[Owen Taylor <otaylor redhat com>]

Federico Mena Quintero <federico ximian com> writes:

> The context of this mail is
> 
> 	http://bugzilla.gnome.org/show_bug.cgi?id=86540
> 
> Basically, there are machines where the maximum visual depth is 8, the
> system visual is 8-bit pseudocolor, but there are 8-bit truecolor
> visuals available.
> 
> Here, gdk_rgb_score_visual() would give the same score to the 8-bit
> pseudocolor and 8-bit truecolor visuals, but it gives preference to the
> system visual and thus picks the 8-bit pseudocolor one.  It allocates
> its 6x6x6 color cube, leaving few colormap entries available for
> non-GNOME applications, and so causing the usual sort of problems when
> you run out of colormap entries.
> 
> Since GdkRGB would be allocating a color cube, anyways, doesn't it make
> sense for it to give preference to truecolor visuals even if they are 8
> bits deep?  The 8-bit truecolor visual will have a color cube of its
> own.  GNOME would run in the truecolor visual, leaving the pseudocolor
> system visual nearly empty for use by other apps.

What worries me is that I've seen XFree86 advertising true/static-color
visuals along with the 8bpp display, but implemented them via colormap
switching. 

In fact, I think this is normal for XFree86 in 8bpp.

Are we sure that the same thing isn't true of this display? 

Regards,
                                        Owen