Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site pluto.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!decwrl!amdcad!amd!pesnta!phri!pluto!warren
From: warren@pluto.UUCP (Warren Burstein)
Newsgroups: net.emacs
Subject: Re: Trouble making xemacs
Message-ID: <302@pluto.UUCP>
Date: Wed, 15-Jan-86 00:21:48 EST
Article-I.D.: pluto.302
Posted: Wed Jan 15 00:21:48 1986
Date-Received: Fri, 17-Jan-86 06:34:34 EST
References: <299@pluto.UUCP>
Organization: The Mayhem Institute
Lines: 32
Keywords: GNU
Summary: I fixed it myself The problem is only on Pyramids


A few days ago I posted a problem I was having with GNU Emacs on a
Pyramid, the gist of it is that "nil" had come unbound.  Well I found
it, the compiler doesn't properly compile a line in init_obarray in
lread.c.  The old line was

  /* Intern nil in the obarray */
  XVECTOR (Vobarray)->contents[hash_string("nil", 3) % OBARRAY_SIZE] = Qnil;

The code generated did not store Qnil into the "contents" array.  I
broke the expression up and it works.  You have to declare:

  int hash;
  Lisp_Object *ptr;

And replace the line with:
  /* Intern nil in the obarray */
  hash = hash_string("nil", 3) % OBARRAY_SIZE;
  ptr = &XVECTOR (Vobarray)->contents[hash];
  *ptr = Qnil;

Anyone who wants to know more about the gory details of the assembler
code can contact me by mail.

We are running OSX 3.2.1

Of course all the source code in this message is copywrited by RMS so
don't go selling it.
-- 

The Maxwell R. Mayhem Institute for Quandary Requiem and Maternal Sciamachy
Accept no substitutes.