Message-Id:	<199904200845.KAA20877@ifm.liu.se>
Date:	Tue, 20 Apr 1999 10:45:27 +0200 (METDST)
From:	Pawel Salek <pawsa AT ifm DOT liu DOT se>
Subject:	Re: List archive?
To:	pgcc AT delorie DOT com
In-Reply-To:	<m10ZTjK-0000GWC@chkw386.ch.pwr.wroc.pl>
MIME-Version:	1.0
Reply-To:	pgcc AT delorie DOT com
X-Mailing-List:	pgcc AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On 20 Apr, Krzysztof Strasburger wrote:
> So, why it works for me? I repeat: try to compile it with the
> _old_ gcc (2.7.2.3) and don't build stagex compilers. This version
> of pgcc has a very odd fast-math bug fixed. I'm really happy that
> it appeared (compiler, not bug;).

IMHO, a compiler that cannot compile itself does not deserve to be
called compiler. (Sorry, developers, I appreciate your hard work but
let's say the sad truth straight in the eyes).

I would not trust a compiler only because it compiles a particular
program. One can try to compile a more complex program and wonder later
if the number one gets are true results or just a garbage.

"Compilers are buggy as long as one cannot prove they are OK"

/Pawel
-- 
Pawel Salek (pawsa AT ifm DOT liu DOT se)      WWW: http://www.ifm.liu.se/~pawsa/ 
Computational Physics Div., Linkoping, Sweden      phone +46 13 281268

- Raw text -