Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!husc6!seismo!mimsy!chris
From: chris@mimsy.UUCP
Newsgroups: comp.lang.c
Subject: Re: short circuit evaluation
Message-ID: <5388@mimsy.UUCP>
Date: Wed, 11-Feb-87 23:34:27 EST
Article-I.D.: mimsy.5388
Posted: Wed Feb 11 23:34:27 1987
Date-Received: Thu, 12-Feb-87 23:29:51 EST
References: <4391@brl-adm.ARPA>
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
Lines: 21
Summary: Bad example; try again

Concerning guarantees as to evalutions, in article <4391@brl-adm.ARPA>
I wrote:
>	a = 0 * *p++ * *q++;
>
>Should we also say that the memory references in the third expression
>must occur? ... if so, must the compiler also do the multiply?  (It might
>overflow, which could be important.)

Bad example; the compiler is certainly allowed (0 * *p++) giving
0 giving * *q++ giving 0, which does not overflow.  Suppose instead
we have

	double *p, r;
	...
	r = *p++ / 1.0;

Must this do the divide, possibly causing (e.g.) a reserved operand
fault on a Vax due to an invalid float value at *p?
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7690)
UUCP:	seismo!mimsy!chris	ARPA/CSNet:	chris@mimsy.umd.edu