annala@neuro.usc.edu (A J Annala) (05/07/89)
The San Diego Supercomputer Center has just installed X11R2 on it's Cray X-MP/48 computer system. At USC I have a color Sun-III/160 running the X11R3 maintained by the campus computer center. I seem to be having a few small problems getting clients on the cray to communicate with the server on my sun. So I was hoping someone in the comp.windows.x community might be able to answer a few questions: o Are there any inherent incompatibilities between X11R2 clients and X11R3 servers? If so, what kind of problems will I encounter? Do you have any software or operating procedure suggestions to solve these problems? o Does anyone else have experience running X11 in this configuration? o Where could I get a copy of the sources for X11R2 if I need to use it? o Does anyone know when a Cray based X11R3 will become available? o If I obtain an official MIT distribution kit for X11R3 via anonymous ftp over the INTERNET, where can I get a full set of prior maintenance patches and where can I continue to obtain future maintenance patches? o Does anyone maintain (is it possible to maintain) either sources or compiled object code or linked executable code in a fully up-to-date version that could be copied and installed on my sun whenever update is required? This would benefit people who need to get on with their research in some area other than computer science but want to use the most current tools to accomplish their research. o What are the Purdue speedups? Where does one obtain prior and current copies of these patches? Are up-to-date source, compiled, or linked code availanle for anonymous ftp? Once again, this would reduce the amount of time non computer scientists would need to spend doing software work. o Does anyone have image processing and/or signal processing packages available for X windows on the Sun? Are these packages sufficiently portable to be moved to the SDSC Cray? Thanks, AJ Annala
rws@EXPO.LCS.MIT.EDU (05/08/89)
I seem to be having a
few small problems getting clients on the cray to communicate with the
server on my sun.
There are lots of 32-bit-isms and other wonders that cause problems on Crays
in both the R2 and (to a lesser extent) R3 sources. Cray has been dutifully
fixing them and letting us know about them, for which we are grateful, and
R3 bug fixes will be incorporated into R4. If you're trying to run a vanilla
MIT R2 on the Cray, good luck. If you got it from Cray, you should be in much
better shape.
If I obtain an official MIT distribution kit for X11R3 via anonymous
ftp over the INTERNET, where can I get a full set of prior maintenance
patches and where can I continue to obtain future maintenance patches?
Patches can be obtained via anonymous ftp to expo.lcs.mit.edu, in
/pub/R3/fixes/.milliken@bbn.com (Walter Milliken) (05/08/89)
In article <17068@usc.edu>, annala@neuro (A J Annala) writes: >o Are there any inherent incompatibilities between X11R2 clients and > X11R3 servers? If so, what kind of problems will I encounter? Do > you have any software or operating procedure suggestions to solve > these problems? > >o Does anyone else have experience running X11 in this configuration? I've been running a similar configuration, using X11R2 clients on a BBN Butterfly GP1000 and the MIT X11R3 Sun server on a 3/50 under SunOS 3.5. I haven't had any major problems with this, just the cursor glitching problem endemic to X11R2 xterm running with the X11R3 server. (Apparently caused by the changed font metrics in R3.) In testing our own X software, we ran into a couple minor "gotcha's" related to the font changes: one program was using font metrics in a bogus way, and was terribly confused when the X11R3 metrics differed from its (wrong) expectations. Another program used a built-in font name without allowing it to be changed. I did run into some problems with the R3 server behaving slightly differently than the R2 one -- one program of mine that mapped many windows worked fine with R2, but would get stomped on by the R3 server (it closed the connection, apparently because the output queue to the client overflowed with expose events). All the other programs I know of just worked. >Thanks, AJ Annala ---Walter