I think this is really an Emacs problem, but Slime is the first Emacs package I've used where problems like this have come up, so I'm asking here first.
I'm running XEmacs 21.4.15 on OS X 10.3.8. I just updated my Slime to the newest, shiniest CVS version (from 1.0 or so), only to find out that it won't work with my XEmacs. Near as I can tell, it's because my XEmacs isn't MULE-ized ({make,find}-coding-system do not exist) and Slime really wants MULE or some semblance thereof to operate. I assume then that Slime does not support non-MULE emacsen...is this a correct assumption? Or do I just need a newer XEmacs that features (some) support for coding systems in non-MULE builds? I see from looking at the archives that people seem to be using Slime CVS in XEmacs, but it's not clear whether those are MULE builds or not.
Thanks,
Nathan Froyd froydnj@cs.rice.edu writes:
I'm running XEmacs 21.4.15 on OS X 10.3.8. I just updated my Slime to the newest, shiniest CVS version (from 1.0 or so), only to find out that it won't work with my XEmacs. Near as I can tell, it's because my XEmacs isn't MULE-ized ({make,find}-coding-system do not exist) and Slime really wants MULE or some semblance thereof to operate. I assume then that Slime does not support non-MULE emacsen...is this a correct assumption?
The assumption is correct, but it's not much work to remove any dependency on those functions. It should suffice, to provide dummy functions for find-coding-system and check-coding-system which return true for 'binary, and maybe a (if (fboundp 'set-process-coding-system)) in slime-open-stream-to-lisp.
I can't try a non-mule version myself, because the Debian "nomule" versions still have find-coding-system. Also, if I configure CVS XEmacs with --disable-mule, I still get find-coding-system as a built-in function.
Helmut.