nanogui: Thread: mwin 0.85

Brad,
	Well, I think what I will try to accomplish tonight is making
changes so that at the very least, microwindows will never send
invalid values to the fb driver.  The simplest way to achieve this
is to convert all user colors to RGB (and still keep defines for the 
palette ones, should someone want them) and then add
support for the 2, 4, 8 bpp palettes, all linked in, rather than separately
as now.
	We should then run on this (since we can't now) and then
look at the slow areas and determine what we need to do to speed them
up.

	I'll get you a tarball tonight, providing my mail interface works,
which it didn't last night.  then perhaps you can build another kernel, which I'll
download tomorrow and try.

Greg

----- Original Message -----
From: Greg Haerr ####@####.####
To: 'Bradley D. LaRonde' ####@####.####
Cc: ####@####.####
Sent: Thursday, October 14, 1999 5:49 PM
Subject: mwin 0.85


> Brad,
> Well, I think what I will try to accomplish tonight is making
> changes so that at the very least, microwindows will never send
> invalid values to the fb driver.  The simplest way to achieve this
> is to convert all user colors to RGB (and still keep defines for the
> palette ones, should someone want them)

Don't we need to junk those palette ones?  They are untrustworthy for static
pseudocolor.  If applications want to specify palette colors, they'll need
to check first with the system palette as to what is available.


> and then add
> support for the 2, 4, 8 bpp palettes, all linked in, rather than
separately
> as now.

Hmmm... that kinda depends on the answer to my previous question.


> We should then run on this (since we can't now) and then
> look at the slow areas and determine what we need to do to speed them
> up.

OK.


> I'll get you a tarball tonight, providing my mail interface works,
> which it didn't last night.  then perhaps you can build another kernel,
which I'll
> download tomorrow and try.

OK.


Regards,
Brad

: Don't we need to junk those palette ones?  They are untrustworthy for static
: pseudocolor.  If applications want to specify palette colors, they'll need
: to check first with the system palette as to what is available.

Yep