Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site ism780c.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!bellcore!decvax!hplabs!sdcrdcf!ism780c!tim
From: tim@ism780c.UUCP (Tim Smith)
Newsgroups: net.lang.c
Subject: Re: A simple non-portable expression tha
Message-ID: <1709@ism780c.UUCP>
Date: Tue, 15-Apr-86 16:21:30 EST
Article-I.D.: ism780c.1709
Posted: Tue Apr 15 16:21:30 1986
Date-Received: Thu, 17-Apr-86 05:29:12 EST
References: <1368@ism780c.UUCP> <32700003@siemens.UUCP>
Reply-To: tim@ism780c.UUCP (Tim Smith)
Organization: Interactive Systems Corp., Santa Monica, CA
Lines: 24

In article <32700003@siemens.UUCP> jrv@siemens.UUCP writes:
>>If R and L are longs, and i is an int, then the expression
>>
>>	L = R + 2 + i;
>>
>>is non-portable.  Lint, however, does not complain about this.
>
>I must be missing something. Why is this non-portable?
>

Since + is associative and commutative, the compiler is free to
re-order the expression.  If you are on a machine where and int
is shorter than a long, and if the compiler decided to do the
2+i first, you may get an overflow, whereas if the compiler
does the R+2 first there is no problem.

To make this expression portable, it needs to be

	L = R + 2 + (int)i;
or
	L = R + 2L + i;

-- 
Tim Smith       sdcrdcf!ism780c!tim || ima!ism780!tim || ihnp4!cithep!tim