[Xquartz-dev] 2.4.0_rc2

Jeremy Huddleston jeremyhu at apple.com
Mon Aug 10 17:50:45 PDT 2009


On Aug 10, 2009, at 17:39, Jack Howarth wrote:

> Jeremy,
>   The rasmol sources have a line...
>
> # USER DEFINITIONS
> #
> #  At this point, you may define the cpp macro PIXELDEPTH as 8, 16  
> or 32
> #define PIXELDEPTH 32
>
> What I find is that with the x86_64 binaries, the rasmol binary built
> with PIXELDEPTH 32 fails with...
>
> Renderer Error: Unable to allocate frame buffer!

I'm guessing it's looking for a particular visual and not finding a  
match... I don't have any clue why this would work on 32bit and not  
64bit builds... they should return the same list of visuals in both  
cases (and from my test with 'glxinfo', I can tell that is the case...)

> whereas for the rasmol binary built with PIXELDEPTH 8 or 16, I get...
>
> No suitable display detected!
>
> This occurs on the Radeon X1600. On the Radeon HD2600, these binaries
> run fine when built at 8, 16 or 32 PIXELDEPTH but the rasmol program
> always reports [24-bit version]. Can you remind me how to debug this
> so that when I run rasmol it will report what visuals are attempted to
> be used?

This will show you helpful environment variables
$ strings /usr/X11/lib/libGL.dylib | grep LIBGL

so do:
export LIBGL_DIAGNOSTIC=1
export LIBGL_DUMP_VISUALID=1

--Jeremy


More information about the Xquartz-dev mailing list