Mailing-List: contact cygwin-help@cygwin.com; run by ezmlm
Sender: cygwin-owner@cygwin.com
Mail-Followup-To: cygwin@cygwin.com
Message-Id: <4.3.2.7.2.20040104123827.02e10450@mailhost.esva.net>
Date: Sun, 04 Jan 2004 13:18:27 -0500
To: "Gerrit @ cygwin" <cygwin@cygwin.com>
From: Beman Dawes <bdawes@acm.org>
Subject: Re: Bug in fstream code and gcc-2 package?
Cc: "John Maddock" <john@johnmaddock.co.uk>
In-Reply-To: <1786768525.20040104122351@familiehaase.de>
References: <4.3.2.7.2.20040103203307.02db8880@mailhost.esva.net> <4.3.2.7.2.20040103092732.02db6dd0@mailhost.esva.net> <01e101c3d1f1$9bc8d6a0$29750252@fuji> <4.3.2.7.2.20040103092732.02db6dd0@mailhost.esva.net> <4.3.2.7.2.20040103203307.02db8880@mailhost.esva.net>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"; format=flowed

At 06:23 AM 1/4/2004, Gerrit P. Haase wrote:

 >Hallo Beman,
 >
 >Am Sonntag, 4. Januar 2004 um 02:55 schriebst du:
 >
 >>>$ g++ -mno-cygwin fstream.cpp
 >> In file included from
 >> /usr/local/include/c++/3.3.1/bits/locale_facets.h:166,
 >
 >The default location for Cygwin & MinGW files distributed by the
 >Cygwin project is /usr but not /usr/local, I suggest that you try to
 >setup a new Cygwin tree (after backing up your old tree).  Somehow
 >this looks wrong to me.  Possibly some old headers around there under
 >/usr/local?

Not just old headers - there is a full installation of gcc 3.3.1 there.

OK, I renamed /cygwin to /old-cygwin, and then did a complete cygwin 
reinstall.

That clears the problem with fstream.cpp!

I think I know what happened - at some point after gcc 3.3.1 became 
available, I downloaded the source and built it myself because 3.3.1 wasn't 
available from cygwin yet. That must have been what created the /usr/local 
copy of stuff. Then later when gcc from cygwin was installed, the 
/usr/local gcc stuff wasn't removed, and it somehow interferes with the 
cygwin supplied version.

I'll run a rebuild of the full Boost regression tests to see what happens 
in real use...

While waiting for that to finish, I'd like to thank you for the prompt and 
helpful response. I occasionally report compiler errors to Comeau, Intel, 
Metrowerks, and Microsoft, and the experience ranges anywhere from 
pleasurable to frustrating to anger producing. Cygwin is high on the 
pleasurable end of the scale.

OK, the tests are finished. Good news! Eleven additional Boost regression 
tests now pass and there are no new failures. The only other difference is 
that the version is being reported as GNU C++ version 3.3.1 (cygming 
special) rather than just GNU C++ version 3.3.1. I assume that is to be 
expected.

Thanks!

--Beman Dawes 


--
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/