Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
To: cygwin AT cygwin DOT com
From: "Hans Horn" <hannes AT 2horns DOT com>
Subject: Re: [ANNOUNCEMENT] Release candidate: gcc-3.4.1-1
Date: Fri, 27 Aug 2004 07:57:24 -0700
Lines: 39
Message-ID: <cgni4m$f8c$1@sea.gmane.org>
References: <200408241633 DOT i7OGXSf02448 AT esds DOT vss DOT fsi DOT com> <Pine DOT CYG DOT 4 DOT 58 DOT 0408241426090 DOT 328 AT fordpc DOT vss DOT fsi DOT com> <cggj9t$gs7$1 AT sea DOT gmane DOT org> <cgitpl$5p5$1 AT sea DOT gmane DOT org> <641907194700 DOT 20040826122845 AT familiehaase DOT de>
X-Complaints-To: usenet AT sea DOT gmane DOT org
X-Gmane-NNTP-Posting-Host: adsl-63-201-90-235.dsl.snfc21.pacbell.net
X-Newsreader: Microsoft Outlook Express 6.00.2900.2180
X-RFC2646: Format=Flowed; Original
X-IsSubscribed: yes

Gerrit,

The problem goes away if I either

leave the -ffast-math flag out (and remove the flags that are implicitly set
by -fmath-math under the hood, i.e. -funsafe-math-optimizations
and -fno-trapping-math)

or leave the -mfpmath=sse flag out

or turn optimization off all-together (replace -On with -g)

Hope this narrows things down a bit!

H.


Gerrit P. Haase wrote:
> Hans wrote:
>
>> Just to follow up on my earlier posting!
>
>> Could it be that I'm picking a stale g77 runtime lib?
>
> I don't think so, looks like a bug in the optimizer.
>
>> The problem doesn't happen if I replace -O6 with -g.
>> It ALSO goes away (with -O6) if I call fabs() instead of abs().
>> fabs() however is a C function, which I don't want rely on!
>
>
> There is a bug with optimization in 3.3.x.  Maybe there are more bugs
> in the optimizer, please try to figure out which flag triggers the
> error.
>
>> I've attached the output of cygcheck -s -v -r, if that's of any help.





--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/