tran@umn-cs.UUCP (03/06/86)
I am having difficulties installing GNU Emacs 17.49 on a vax-750 running System 5.0 version 2. Make complained that the size of structure nlist (or something like that) in file FNS.C, line 1040 is undefined. I then explicitly included <nlist.h> inside FNS.C; this seemed to work fine. But then the C compiler gave a warning that malloc() does not return the type ptr but integer, when compiling ALLOC.C. I just ignored it. The next error is GETPAGESIZE() was undefined in ALLOCA.C. I replaced it with NBPC after including <sys/param.h>. (not sure what I'm doing here) The real difficulty came when TEMACS started loading standard lisp packages and called (garbage-collect). It gave the message 'Wrong type of argument: stringp, 1329' and hang until I typed control-D. I have tried to fix it, but the source code looks formidable. Could somebody please help ? The other things I have tried are using different s-usg5*.h; casting pointer type to malloc, and deleting (garbage-collect) from loadup.el. Still didn't work. PS: I also installed it on a SUN running 4.2BSD, successfully. DOCTOR is pretty neat. -Nicholas Tran
hrp@cray.UUCP (Hal Peterson) (03/09/86)
> > I am having difficulties installing GNU Emacs 17.49 on a vax-750 running > System 5.0 version 2. I've been having identical difficulties with 17.49 on a VAX 785 running System V.2. I used the same workarounds. > The real difficulty came when TEMACS started loading standard lisp packages > and called (garbage-collect). It gave the message 'Wrong type of argument: > stringp, 1329' and hang until I typed control-D. I got around this by commenting out all of the (garbage-collect) lines in loadup.el and putting (setq gc-cons-threshold (* 2 gc-cons-threshold)) at the beginning of loadup.el. Note that that's a WORKAROUND, NOT A FIX. Anyway, once that was done, temacs ran to completion and dumped to xemacs. The bad news is that this xemacs now says "Fatal error." and dumps core with a bus error when I try to run it. I have no idea why, or how to find out why. Has anybody out there done this already? Or does somebody have a guess on what's broken? My time for this sort of work is limited, and I'm not eager to reinvent the wheel. > -Nicholas Tran -Hal Peterson -- Hal Peterson / Cray Research / 1440 Northland Dr. / Mendota Hts, MN 55120 UUCP: ihnp4!cray!hrp phone: (612) 681-3085