Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!cs.utexas.edu!wuarchive!udel!princeton!notecnirp!nfs
From: nfs@notecnirp.Princeton.EDU (Norbert Schlenker)
Newsgroups: comp.os.minix
Subject: Re: NULL
Keywords: NULL 0 ANSI
Message-ID: <21229@princeton.Princeton.EDU>
Date: 13 Nov 89 03:05:08 GMT
References: <4455@ast.cs.vu.nl> <1347@gould.doc.ic.ac.uk>
Sender: news@princeton.Princeton.EDU
Reply-To: nfs@notecnirp.UUCP (Norbert Schlenker)
Organization: Dept. of Computer Science, Princeton University
Lines: 21

In article <1347@gould.doc.ic.ac.uk> dds@cc.ic.ac.uk (Diomidis Spinellis) writes:
> appeals to authority omitted
>...
>3. Will defining NULL as ((void *)0) enable some WRONG programs to
>   successfuly run on some architectures and C implementations?	[yes]
>...
>Yes.  In some architectures and some C implementations WRONG programs
>will work.  Here is an example for this:
>...
> example of broken code that works correctly if NULL is ((void *) 0) omitted

The argument is specious.  It is not hard to come up with broken code, similar
to that of the article, which works when NULL is defined as 0 and not when it
is defined as ((void *) 0).

All things considered, I think it will be more portable to define NULL as a
simple 0.  I don't think that appeals to Harbison and Steele are germane,
and the draft ANSI standard leaves it (purposely) vague.  It's just simpler
to use 0.

Norbert