Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!cs.utexas.edu!samsung!dali!uakari.primate.wisc.edu!zaphod.mps.ohio-state.edu!mips!pacbell.com!pacbell!att!dptg!ulysses!andante!alice!andrew
From: andrew@alice.UUCP (Andrew Hume)
Newsgroups: comp.lang.c
Subject: Re: ANSI C free() NULL test "was free (NULL)";
Summary: not the point
Message-ID: <10850@alice.UUCP>
Date: 22 May 90 14:11:42 GMT
References: <1194@wet.UUCP> <2855@mica6.UUCP> <751@mwtech.UUCP> <2977@mica6.UUCP> <7@agtoa.UUCP>
Organization: AT&T Bell Laboratories, Murray Hill NJ
Lines: 9



the badness of allowing free(NULL) is not so much that it encourages sloppy
programming but rather its tie with the bad decision to make a
special case out of malloc(0) which ought to have meant what every manual page
until recently said it meant: a pointer to at least 0 bytes of memory.
no need for a calculus of zero sized objects; just follow the man page.

still, who am i to complain; at least the committee got noalias right.