Path: utzoo!mnetor!uunet!husc6!tut.cis.ohio-state.edu!karl
From: karl@tut.cis.ohio-state.edu (Karl Kleinpaste)
Newsgroups: comp.emacs
Subject: Re: Problem bringing up GNU Emacs on sbe20
Message-ID: <11058@tut.cis.ohio-state.edu>
Date: 20 Apr 88 02:40:26 GMT
References: <215@richp1.UUCP>
Organization: OSU
Lines: 30
In-reply-to: craig@richp1.UUCP's message of 19 Apr 88 17:49:52 GMT


craig@richp1.UUCP writes:
   I'm trying to bring up GNU Emacs on an sbe20 (68020 based SYSVR2 box),
   and have been somewhat unsuccessful.  Once temacs is made, and
   starts running the message:
	   ^N^?^?^?|Ho: is not an identifier
   comes up, causing substantial grief.

That looks like an error from shell, not the program itself, which in
turn would suggest that the resulting binary isn't even a valid binary
in the opinion of your kernel.  That's a neat trick if you used your
normal C compiler.

   Since the pre-crt0 /lib/crt0 combination doesn't work (Termination
   code 139?),

Termination code 139 is the exit status noted by make(1) when temacs
ran.  Look up wait(2) and observe that a non-zero low byte means
"death by signal," and that if the 0200 bit is set, you got a core
dump as it went down in flames.  139 decimal is 0213 octal, which is
signal 11, SIGSEGV.

My first, standard, stock suggestion for anyone porting to Yet Another
SysV Port is
	#define NO_REMAP
in the appropriate file (config.h? m-<whatever>.h?) and recompile from
scratch.  That will avoid a whole herd of possibilities, and for those
SysV implementations that do VM in The Proper Manner, it won't cost
your system any grief, either.

--Karl