Path: utzoo!utgpu!watmath!att!occrsh!uokmax!unmvax!aplcen!haven!ames!excelan!ames!ncar!gatech!hubcap!ncrcae!ncr-sd!hp-sdd!ucsdhub!sdcsvax!beowulf!mattson
From: mattson@beowulf.ucsd.edu (Jim Mattson)
Newsgroups: comp.sources.games.bugs
Subject: Problems with nethack 3.4
Message-ID: <7206@sdcsvax.UCSD.Edu>
Date: 6 Oct 89 18:19:23 GMT
Sender: nobody@sdcsvax.UCSD.Edu
Reply-To: mattson@beowulf.UCSD.EDU (Jim Mattson)
Distribution: na
Organization: CSE Dept., U.C. San Diego
Lines: 21

I've already seen several postings about problems with this version of
nethack, but most of these posters have at least managed to compile the
program, and I can't even get that far.  I have had several problems on a
machine that compiled 3.3 without even the slightest hiccup (Sun 3 running
SunOS 3.5):
1. Make complains about predecessor loops, blasts away some of the 
   dependencies, and ends up doing things in the wrong order.
2. NUMMONS undefined in you.h (Okay, I saw the posting today that fixes
   this).
3. Here's the real big one:  My patches added a field to the objclass struct,
   but there are no patches to objects.c, where all of the static
   initialization of objects[] takes place.  As a result, o_let ends up
   at the wrong field offset, and makedefs -o complains about probability
   errors for several ridiculous symbol types before dumping core.
   So, what gives?  Did I miss some patches to object.c, or what?

--jim

	Internet: mattson%cs@ucsd.edu
	Bitnet:	jmattson@ucsd
	UUCP:	{ucbvax|rutgers|bellcore}!ucsd!cs!mattson