To get information about your X11 color depth
use:

 xdpyinfo
---------------------------------------------------------------
Date: Tue, 11 Apr 2000 23:06:16 -0400               
From: Carl Anderson <candrsn@mindspring.com>
This seems good enough to rollout ...               
                        
The problem you were experiencing was from a bug in 
XFree86. I have coded around that and not been able to                          
reproduce the dead looping behaviour.               
                        
I appreciate your desire to maintain backward compatibility.                    
                        
As much as possible I have coded this to be generic and to maintain             
the previous behaviour. 
                        
Of course where I have converted reliance on DefaultXXX to                      
parameters, I may have introduced errors.           
                        
I tested this code as much as possible for me       
                        
under Linux - i386 (using an ATI Mach64 card, and an S3 Virge Card)             
  startx -bpp 8  (after editing /etc/X11/XF86config)                          
        I changed the desired visual in Graph_Set.c to each of                  
        TrueColor, DirectColor, PseudoColor, StaticColor,                       
        GrayColor, and StaticGray      
  startx -bpp 16 (after editing /etc/X11/XF86config)
        my visual is TrueColor Only                 
  startx -bpp 24 (after editing /etc/X11/XF86config)
        TrueColor only  
                        
under Linux - sparc (using a cgsix card)          
  startx -bpp 8  (after editing /etc/X11/XF86config)                          
        I changed the desired visual in Graph_Set.c to each of                  
        TrueColor, DirectColor, PseudoColor, StaticColor,                       
        GrayColor, and StaticGray                   
                        
under Solaris 2.7 - sparc and utlrasparc            
  /usr/openwin/bin/openwin  (using a cgsix card in 8bit mode)                 
        I changed the desired visual in Graph_Set.c to each of                  
        TrueColor, DirectColor, PseudoColor, StaticColor,                       
        GrayColor, and StaticGray                 
                      
  /usr/openwin/bin/openwin (using a pgx24 card in 8 bit mode)                   
        I changed the desired visual in Graph_Set.c to each of                  
        TrueColor, DirectColor, PseudoColor, StaticColor,    
        GrayColor, and StaticGray                 
                        
  /usr/openwin/bin/openwin (using a ffb card in 8/24 bit mode)                  
        I changed the desired visual in Graph_Set.c to each of                  
        TrueColor, DirectColor, PseudoColor, StaticColor,                       
        GrayColor, and StaticGray                 
                        
under each test the driver ran,                     
        Truecolor and PsuedoColor ran correctly (with correct colors)           
        DirectColor,StaticColor,StaticGray ran correctly (with alien colors)    
                        
I have not seen a Display which was limited to StaticColor or StaticGray        
or DirectColor without either PsuedoColor or TrueColor available.          
                        
As this patch fixes a deficiency in GRASS for XFree86                           
and the limitations of teh patch don't seem likely to be exposed (IMHO), 
BTW  without this patch I cannot run GRASS  under        
        startx -bpp 8 on the XFree86 Linux Machines 
I additionally have trouble with Solaris            
        when running 8 bit visuals.                 
                        
should you reqiure a dump of xdpyinfo for each of these card i tested           
it on I can provide them.                           
                        
The differences between this patch and the one I posted to Grass-List           
is better error trapping and better informational messages.                     
                        
of course the new parameter                         
   nlev=-1   to force TrueColor, if possible        
needs to be advertised.      

Carl Anderson
