josh@mit-vax.LCS.MIT.EDU (Joshua Marantz) (09/19/87)
I have successfully installed X11 on a Sun 3/160C system. I had to tweak a bunch of things (other than the distributed patch), but it works. Some notes: 1. gnuplot does not compile or install on the sun. Be forewarned. 2. You have to rebuild the makefiles with imake. The distributed versions are for the Vax. I modified util/imake/imake.tmpl to only work with suns, cause I wasn't confident the #ifdef vax, etc. would work. Do all this BEFORE you compile anything. If you have the string Vax.macros in any of your Makefiles, then you've done something wrong. 3. If you have diskspace problems, you can delete the .o files after the library (.a) has been built. Of course, if you want to make a change to one module, the stupid Makefile will recompile everything. Ideally, the Makefile should compare c-file dates to library member insertion dates, but I'm not sure you can do this in Unix make. A ridiculous workaround might be to leave a dummy .o file with the same timestamp as the real .o that you deleted. I'm sure someone would get burned by this practice, though. 4. Few of the clients work correctly on the Sun, but wm is reasonable. 5. Everything is slow: The mouse is too slow -- Window dragging is painful. xterm refreshing is incredibly slow. Far worse than 1200 baud. I compiled with -DSUN_WINDOWS -DDEBUG. Was this correct? Should I run xinit with any special options? I don't think I have a GP. X10R4 ran orders of magnitude faster on this machine. 7. Haven't tried recompiling GNU Emacs yet. Unfortunately, the old version won't even work inside an xterm window, so I'm screwed until I do. I guess I won't delete my SunView binaries yet. I probably can rlogin inside an xterm window and then run emacs. I bet the network overhead won't be noticable next to the 300-baud xterm. -Joshua Marantz Viewlogic Systems, Inc.