Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!usc!zaphod.mps.ohio-state.edu!rpi!crdgw1!uunet!brunix!doorknob!dfs
From: dfs@cs.brown.edu (David Sklar)
Newsgroups: comp.sys.mac.programmer
Subject: Access to hardware framebuffer (Color QD)
Message-ID: <DFS.90Aug20184133@dooby.cs.brown.edu>
Date: 20 Aug 90 22:41:33 GMT
Sender: news@brunix.UUCP
Distribution: comp
Organization: Department of Computer Science, Brown University
Lines: 34

I am attempting to use the MacII to run interactive programs that were written
for framebuffers before the days of palettes, virtual LUTs, etc.  These
programs expect the pixel values they draw to *not* be modified before being
placed in the hardware framebuffer.  Thus, I am bypassing the Palette Manager
and trying to get as close to the hardware as possible.  I want to be able to
set the hardware color table with extreme precision: to set LUT entry 0 and
know that entry 0 (exactly) of the hardware LUT is being set.  Likewise, to set
a pixel in the screen's pixmap to value X and know that exactly X is being
placed in the hardware framebuffer.

Let's say I want to draw a line in color 0.  I have reason to believe (but no
proof) that it's not enough for me to simply:

	1) Set the port's "fgColor" field to 0 
		(manually, not using a QD routine)
	2) Call QuickDraw to draw the line.

Of course, there's no way to for me to verify that a 0 was not placed in the
hardware FB.  But I believe my pixel values are getting mapped to different
numbers, because XOR'ing is not working as one would expect.  For example, 
	1) I specified a white color for LUT entry 0
	2) I specified a black color for LUT entry 1
	3) I drew an image using only fgColors 0 and 1
		(always manually setting the fgColor field)
	4) I set the transfer mode to XOR
	5) I draw a line using fgColor 1
		THE RESULT was *not* an inversion of all pixels touched
			by the line.  It seems the numbers I specify are
			being mapped to something else before they are
			placed in the hardware FB.

Any help in furthering my understanding of this under-documented method for
using Color QD would be appreciated!