Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!zaphod.mps.ohio-state.edu!sdd.hp.com!ucsd!ucbvax!bloom-beacon!lightning.mcrcim.mcgill.EDU!mouse
From: mouse@lightning.mcrcim.mcgill.EDU
Newsgroups: comp.windows.x
Subject: Re:  problem with colormap
Message-ID: <9011091424.AA06728@lightning.McRCIM.McGill.EDU>
Date: 9 Nov 90 14:24:15 GMT
Sender: daemon@athena.mit.edu (Mr Background)
Organization: The Internet
Lines: 56

> I was trying to get a ray tracing program that someone else wrote to
> work on my DECStation 5000/200.  The program fails when it attempts
> to display an image and I traced the problem down to a section of
> code that creates a new colormap.  Below is the code fragment that I
> think is causing problems.

[edited for compactness -dM]
> 	i = XMatchVisualInfo(display, theScreen, depth, PseudoColor, &visualinfo);
> 	if (!i) i = XMatchVisualInfo(display, theScreen, depth, TrueColor, &visualinfo);
> 	if (!i) exit(1);
> 	visual = visualinfo.visual;
> 	rootw = RootWindow(display, theScreen);
> 	theGC = DefaultGC(display, theScreen);
> 	cmap = XCreateColormap(display, rootw, visual, AllocAll);
> ...
> 	XStoreColors(display, cmap, carray, CMAPSIZE);

> The first call to XMatchVisualInfo always fails but the second call
> succeeds so apparently the visual class is TrueColor.

(Or rather, no PseudoColor visual is available but a TrueColor visual
is.  On color displays there often (usually?) is no "the" visual
class.)

> This causes the XCreateColormap call to fail with a BadMatch error.
> I tried fixing this problem by passing AllocNone instead of AllocAll
> but this caused a protection error when XStoreColors was called.

The problem is that you simply can't treat PseudoColor and TrueColor
the same.  For PseudoColor you need to

	XCreateColormap(...,AllocAll);
	XStoreColors(...);

whereas for TrueColor you simply

	XCreateColormap(...,AllocNone);

Note though that if the *default* visual is a TrueColor then there is
no need to create a new colormap; you can simply use the default.

> Can anybody offer any insight into this problem and suggest a
> workaround ?

I would guess that the code was developed on a system with PseudoColor
available, so the TrueColor code was untested.  For a workaround, I
would recommend remembering somewhere whether the visual is PseudoColor
or TrueColor and doing the colormap differently, as I sketched above.
Note that you need to take care that your pixel values are correct for
the desired colors in the TrueColor case; you can't assign them
yourself as you can for PseudoColor.

					der Mouse

			old: mcgill-vision!mouse
			new: mouse@larry.mcrcim.mcgill.edu