Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!samsung!spool.mu.edu!uwm.edu!linac!att!ucbvax!bloom-beacon!dont-send-mail-to-path-lines
From: mouse@lightning.mcrcim.mcgill.EDU (der Mouse)
Newsgroups: comp.windows.x
Subject: Re:  Color maps....
Message-ID: <9103280703.AA16051@lightning.McRCIM.McGill.EDU>
Date: 28 Mar 91 07:03:05 GMT
Sender: tytso@athena.mit.edu (Theodore Y. Ts'o)
Organization: The Internet
Lines: 49

> I am very new to the world of X programming so pardon me if I sound
> stupid or that I didn't read the books thorough enough...

This is a relatively frequent question, but you shouldn't have much
trouble with it.  You are already past the first hurdle most people
seem to get stuck on.  Fortunately, your problem is solvable.

> The problem is, when I need to do logics with the colors, like
> xoring, anding etc (e.g. white on blue gives yellow..etc), since I am
> not using the default colorcells, I can't get the correct color for
> the modes.

I don't know what you're doing with 16 colors, when there are only
three color channels, if you expect to work with primary colors this
way, but that's your business....

> It's not as simple as blue (0001) on white (1111) xor gives yellow
> (1110) because there's no way to know what color cells are allocated
> (or is there?)

Well, you know what cells are allocated because they're returned by the
allocation call.

> and even if we know, there's no systemetic scheme to do the logics.

Right.  You have to get lucky and get cells that happen to have the
relationships you want.

Now, what's the solution I implicitly promised above?  Simply this: you
tell the server to give you cells that are related the way you want.
To do this, you have to allocate read/write cells instead of read-only
cells (minor wart, but more or less unavoidable).  What you actually
want is to have a group of color cells that have all but four bits
equal, then those four bits give you the sixteen combinations you want.
The routine to allocate colormap cells this way is XAllocColorCells;
you would want to call it with nplanes=4 and npixels=1.  If it's
important that all four bits be adjacent (which I gather in your case
it is not), you can set contig True; otherwise, set it False.

Of course, you can't blindly use GXxor even in this case; you need to
take care to avoid clearing the "common" bits.  You can do this by
setting the plane-mask in the GC or by being careful to set the
foreground and background values to something appropriate for the
operation in question, take your pick....

					der Mouse

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