🚀 go-pugleaf

> I didn't have any luck with Xibm-athena either.  brazos.is.rice.edu
> appears to have the server and the fonts, but not the libraries or
> clients.  Maybe the good people at Rice would be willing to share the
> changes they had to make to build the server for RT AIX.  Apparently
> the code needs a lot of tweaking to get past the RT compiler.

Hello all -

My apologies for not having spoken up until now on this subject, but as
the end of semester approaches and I worry about finding a job with my
newly-graduated fiancee I find that there's less time to read news ...

I am the one who compiled the X11R4/R5 distributions available from 
brazos.is.rice.edu.  Yes, it did take a significant amount of tweaking 
to get the code working - Most of the work involved getting the PEX code
revised to do things the non-ANSI way at the flip of a preprocessor macro.  
However, the really tough part of the port was in the server/ddx/ibm subtree.  
As such, I'll note that the changes required to get the server to run are
divided into two sections - the changes to the ibm subtree and the changes
to the rest of the mit code.  I have sent all of the changes made in the 
ibm subtree portion of the code to MIT and they have been folded into John 
Carr's X11R5 port available from demi.mit.edu.  I have not yet generated
a list of patches for the remainder of the distribution.

I've been hesitating to make the remainder of the changes available because
they are not finished yet.  It is true that I have the server working - we're
using it very successfully here at Rice - but I got the thing running in a
3-day hackfest that I'm not particularly proud of.  Your posts have renewed
my interest and I'll be trying to clean up the code as time permits.

So ... for those of you out there who want to use X11R5, feel free to use
the server as supplied in pub/aix_rt on brazos.is.rice.edu.  We have used 
it here to turn our army of RT's into X-terminals, using XDM to display
off of SPARC compute servers.  That is why I did not compile any of the 
clients - we just don't use them.  I would suggest this as the ideal 
configuration for any of you out there using RT's.  They make great X-servers, 
but a single RS/6000 320H or SPARCStation 2 will outperform an army of RT's 
supporting a complete user environment locally.  In addition, the 
configuration and system management costs of keeping an RT useful go to 
zero when they're used this way.

And for those of you out there who have to build X11R5, start with 
John Carr's X11R5 port available from demi.mit.edu.  The remainder of
changes necessary are straightforward, but very numerous.  If you want 
to use my hacked changes to the rest of the mit distribution, just send
me mail and I'll send you a drop of my latest efforts.  If you can wait,
I hope to get my final version completed and posted to comp.sys.ibm.pc.rt 
by May 15.

Again, I apologise for having been silent for so long, but I encourage all
of you to use the X11R5 server available on brazos.is.rice.edu.  If you find
problems with the server, please send me mail and we can work to find a
solution.

Michael Brantley
Office of Networking and Computing Systems
Rice University, Houston, Texas
(brantley@rice.edu)