Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxr!mhuxt!houxm!whuxl!whuxlm!akgua!gatech!seismo!hao!hplabs!sdcrdcf!trwrb!trwrba!cepu!ucsbcsl!frew
From: frew@ucsbcsl.UUCP
Newsgroups: net.micro.att
Subject: 7300 3.0 cc bugs
Message-ID: <395@ucsbcsl.UUCP>
Date: Fri, 28-Mar-86 13:18:55 EST
Article-I.D.: ucsbcsl.395
Posted: Fri Mar 28 13:18:55 1986
Date-Received: Tue, 1-Apr-86 07:27:15 EST
Organization: U.C. Santa Barbara
Lines: 47

Here are a couple of code sequences that make our brand-new PC 7300 Release
3.0 C compiler lose its lunch.  The error messages have been inserted as
comments above the offending lines.  Successful workarounds are enabled by
-DFIX.
---------------------------------------
f(a, b, c)
	double	*a;
	double	*b;
	double	*c;
{
#ifndef	FIX
	/* "b00.c", line 8: compiler error: allocation fails, op STAR */
	*a = *b = *c;
#else
	*b = *c;
	*a = *b;
#endif
}
---------------------------------------
f(p)
#ifndef	FIX
	register double	*p;
#else
	double		*p;
#endif
{
	double		d;

	/* "b01.c", line 11: compiler error: illegal byte address (ZB) */
	*p++ = d;
}
---------------------------------------
There is obviously some gotcha connected with (double *) formal parameters ...

I am posting these in hopes that someone from AT&T is reading this newsgroup.
If anyone out there knows of a better way to email bug reports to AT&T, please
advise.

P.S.:  I would be interested in talking with anyone who is trying to make any
non-trivial use of the 7300's floating point (yes, I realize it's emulated
and not intended for serious number crunching, but it still ought to work,
albeit slowly).

-- 
James Frew	ucbvax!ucsbcsl!frew
Computer Systems Lab., Univ. of Calif., Santa Barbara, CA 93106
(805) 961-2309