Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!olivea!uunet!enterpoop.mit.edu!cambridge-news.cygnus.com!athena.mit.edu!raeburn
From: raeburn@athena.mit.edu (Ken Raeburn)
Newsgroups: comp.os.386bsd.bugs
Subject: Re: gcc 2.3.3 bug?
Message-ID: <RAEBURN.93Mar30122651@cambridge.mit.edu>
Date: 30 Mar 93 17:26:51 GMT
References: <1p7c3mINNdh5@urmel.informatik.rwth-aachen.de>
Organization: Massachusetts Institute of Technology
Lines: 35
NNTP-Posting-Host: cambridge.cygnus.com
In-reply-to: kuku@acds.physik.rwth-aachen.de's message of 29 Mar 1993 17:36:54 GMT
In article <1p7c3mINNdh5@urmel.informatik.rwth-aachen.de> kuku@acds.physik.rwth-aachen.de (Christoph Kukulies) writes:
#include <limits.h>
int i = -2147483648;
main() {
printf("%d %d\n",INT_MIN,i);
}
My gcc2.3.3 issues a warning: integer constant so large that it is unsigned.
Obviously it is in the range of INT_MIN and INT_MAX.
Actually, no, it isn't.
According to ANSI C, the minus sign isn't part of the number, so it
parses as
UNARY_MINUS NUMBER
and the NUMBER is large enough that it becomes unsigned. Then this
large unsigned constant gets negated, which produces another large
unsigned constant.
A proper definition for INT_MIN on most machines would be something
like
#define INT_MIN (- INT_MAX - 1)
since it's a compile-time constant, but is signed.
~~ Ken Raeburn
~~ raeburn@cygnus.com, raeburn@mit.edu
~~ "Your Jedi mind tricks won't work on me, monkey boy!"