Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!samsung!brutus.cs.uiuc.edu!apple!uokmax!occrsh!att!cbnewsm!gregk
From: gregk@cbnewsm.ATT.COM (gregory.p.kochanski)
Newsgroups: comp.lang.c++
Subject: Zortech C++ v2.0
Message-ID: <6376@cbnewsm.ATT.COM>
Date: 9 Nov 89 13:31:51 GMT
References: <5759@spudge.UUCP>
Reply-To: gregk@cbnewsm.ATT.COM (gregory.p.kochanski,mh,)
Organization: AT&T Bell Laboratories
Lines: 38

The following code yeilds the wrong result in zortech C++ 2.0.
Note that it is just straight "C".
(I have excerpted the interesting section from a longer compiler excerciser).
MESS3 should NEVER be printed, and in fact isn't on AT&T C++ 2.0, Microsoft C
v5.1, the native Alliant C compiler, and Zortech 1.07.

#define MESS3 "CPUTEST FAILS ON LOGICALS\n"

main()
{
register int a;
short d, i, *q, **r;
register short int *p;
long c, l, mm;
register unsigned e;
double x;
float yy;
extern short cwords;
static float xx;
static double y;
static long m;
static short int b;

for(a=0;a<2;a++)
	{
	for(b=0;b<2;b++)
		for(d=0;d<2;d++)
			for(c=0;c<2;c++)
				if((a||b)==((!a)&&(!b)) ||
					  (!(c&&d))!=((!c)||(!d)) ||
					(!a)==a || ((a!=c)||a)!=((a!=c)||c))
				write(2, MESS3, sizeof(MESS3)-1);
	}
 exit(0);
}

Greg Kochanski gpk@physics.att.com  Physics for fun and profit.
I may be biased, but this is still true.