RE: [ANNOUNCEMENT] Server Test 36

["Harold Hunt" <huntharo at msu dot edu>]

Holger,

> When I minimize the whole xserver window and raise it again,
> colors may change
> slightly and can be repaired by "drawing" with the cursor. I just now have
> started Netscape with xwin minimized, and after raising it again,
> all window
> colors (xterm, MAGIC) change totally under the moving cursor.

By the time I figured out that fixing this would be difficult I had already
fixed it.  Turns out that the device context returned by BeginPaint () has
to have the palette selected into it and realized on it, in order for the
colors to be correct when redrawing portions of the window that have been
exposed from behind another window.  I'll put together another test release
soon, but maybe not tonight.

> A final feature request now is to put 8bit on top of 32bit, so
> the graphics
> card may stay with 32bit.

It turns out that this is an age old problem that is really difficult to
solve.  I thought I could do it by just initializing a fake PseudoColor
visual when running in TrueColor mode.  Turns out I was wrong.  However,
you're in luck, because the new RandR (resize and rotate) extension by Keith
Packard et al. solves this difficult problem.  RandR provides all visual
classes in all depths to all clients all the time, regardless of the actual
physical display depth.  RandR is not currently enabled in Cygwin/XFree86,
as it appears to still be beta, and some work will have to be done to enable
it.  I'm kinda looking at it now, but you may wish to enable it for me.  I'm
reading the documentation at:
http://www.xfree86.org/~keithp/talks/randr/randr/

You may be wondering why we don't just use Windows palettes in TrueColor
mode to enable PseudoColor.  Well, Windows only allows palettes in 8 bit
color mode, so there you have it.

I'm going to remove the code that created a fake PseudoColor visual when
running in TrueColor mode.  This will have no effect on the real PseudoColor
support that has just been released.

Harold