On Wed, 14 Apr 1999, Jonathan Stone wrote:

> Michael has changes in his R5 tree which make this work much better.

He's sending them to me. Hrm. Lemme think.. How did you handle the TC mfb?
I've just popped in cfbSetupScreen() (8-bit), and used a StaticGray visual
with 1-bit depth. Also, there are apparently 16 and 32 bit versions of the
sfb. Do we support those? Adding the needed code is simple. If we have px
support for 1.4, do we want a half assed, slow but fully working Xserver
for it (I fixed the event/colormap problems while making Xpmax)?

> >I'll fix this once I've gotten some kernel work finished - thanks for
[...]
> Please let me know if you need help.

Ahh.. Have you tried the latest -current? All the rcons/px changes are in
there (I'll give you a list of revsions etc RSN). To get a kernel to
build, I had to #ifdef out htonl et al. in mips locore.S (it's duplicated
in lib/libkern/arch/mips) and #ifdef out .abicalls in byte_swap_?.S,
otherwise ld failed with a message I don't understand. I recall seeing the
same problem about 4 months ago. Should these "solutions" be committed? 

> Also, once we have R6 Xservers, we should try and get them into xsrc
> and tagged onto the release branch. We should drop a note to releng
> letting them know this ugly blemish is about to be fixed, too :). 

I'll put the actual code and diffs against toplevel Imakefile in my
homedir on cvs.netbsd.org RSN.

Andy.