Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!elroy.jpl.nasa.gov!ncar!gatech!bloom-beacon!LARRY.MCRCIM.MCGILL.EDU!mouse
From: mouse@LARRY.MCRCIM.MCGILL.EDU
Newsgroups: comp.windows.x
Subject: Re: Can XAllocColor/XAllocNamedColor fail ?
Message-ID: <9011112253.AA20761@Larry.McRCIM.McGill.EDU>
Date: 11 Nov 90 22:53:19 GMT
Sender: daemon@athena.mit.edu (Mr Background)
Organization: The Internet
Lines: 51

>> XAllocColor finds the closest color supported by the visual, not the
>> closest color in the colormap.  It will fail if the requested color
>> is not in the colormap and the colormap is full.

(That's true for GrayScale, PseudoColor, and DirectColor.  The first
half of the first sentence is true for all six visuals; the second
sentence is vacuously true for the three static visuals.  (But see my
note [$] below.))

> This leads me to a problem Im having, in that what exactly is
> closest, and what is the best way to allocate colors so that you get
> approx. what your looking for, even if the color map is getting
> "full"

It might not be possible.  If someone has nearly filled up the colormap
with a grayscale ramp, so there are only two cells left, and you want
red, green, and blue, there really isn't much you can do: you simply
can't get anything that is even approximately what you're looking for
without creating your own colormap.

Allocating a color on any visual will always, as a first step,
approximate the requested color as closely as possible for that visual.
Taking the resulting approximation as "the requested color", then

	- If the requested color is present in a read-only cell of the
	   colormap, that cell is returned[#];
	- Otherwise, if there is a cell free[%], it is filled in with
	   the requested color and returned;
	- Otherwise, an error is generated.

Note that the first step will always succeed for a static visual,
because the colormap always contains all the colors the visual
supports.

[%] For DirectColor, this may be[$] more complicated than "if there is
    a cell free", because there are three distinct mapping tables
    involved, one for each primary.  I haven't looked at it in detail.

[$] I am not certain XAllocColor works for decomposed visuals
    (DirectColor and TrueColor); I haven't checked in detail.

[#] What happens if more than one read-only cell contains the color is
    not quite as clear, though some one of them should be returned.
    Due to the semantics of allocating colors, this can happen only
    when a server implementor provides a static visual with redundant
    cells in the colormap.

					der Mouse

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