Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT sources DOT redhat DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin/>
List-Post: <mailto:cygwin AT sources DOT redhat DOT com>
List-Help: <mailto:cygwin-help AT sources DOT redhat DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-owner AT sources DOT redhat DOT com
Delivered-To: mailing list cygwin AT sources DOT redhat DOT com
Date: Fri, 20 Oct 2000 13:53:05 -0400
From: Christopher Faylor <cgf AT redhat DOT com>
To: cygwin AT sources DOT redhat DOT com
Subject: Re: Cygwin SEGVs
Message-ID: <20001020135305.C14359@cygnus.com>
Reply-To: cygwin AT sources DOT redhat DOT com
Mail-Followup-To: cygwin AT sources DOT redhat DOT com
References: <20001017120354 DOT A9810 AT cygnus DOT com> <20001017205249 DOT 7782 DOT qmail AT lizard DOT curl DOT com> <20001020053504 DOT 5350 DOT qmail AT lizard DOT curl DOT com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.3.6i
In-Reply-To: <20001020053504.5350.qmail@lizard.curl.com>; from jik@curl.com on Fri, Oct 20, 2000 at 01:35:04AM -0400

On Fri, Oct 20, 2000 at 01:35:04AM -0400, Jonathan Kamens wrote:
>A few days ago, I wrote:
>
>>  >  If you're really motivated, you can build cygwin yourself
>>  
>>  Alas, much easier said than done.  My attempt to build was failing in
>>  all kinds of strange ways, apparently related to configure getting the
>>  wrong answers about the questions it was asking.  The first wrong
>>  answer I noticed in my build log was:
>>  
>>    checking whether gcc -L/scratch/jik/cygwin/build/i686-pc-cygwin/winsup -L/scratch/jik/cygwin/build/i686-pc-cygwin/winsup/cygwin -L/scratch/jik/cygwin/build/i686-pc-cygwin/winsup/w32api/lib -isystem /scratch/jik/cygwin/src/winsup/include -isystem /scratch/jik/cygwin/src/winsup/cygwin/include -isystem /scratch/jik/cygwin/src/winsup/w32api/include -isystem /scratch/jik/cygwin/src/newlib/libc/sys/cygwin -isystem /scratch/jik/cygwin/src/newlib/libc/sys/cygwin32 -B/scratch/jik/cygwin/build/i686-pc-cygwin/newlib/ -isystem /scratch/jik/cygwin/build/i686-pc-cygwin/newlib/targ-include -isystem /scratch/jik/cygwin/src/newlib/libc/include accepts -g... no
>>  
>>  After several hours of trying various things to debug this, I finally
>>  relized that the problem is that many of the tests in the configure
>>  scripts rely on empty output from gcc to indicate that there were no
>>  errors, but gcc was generating the warning "file patch prefix
>>  `/scratch/jik/cygwin/build/i686-pc-cygwin/newlib/' never used" to
>>  stderr, thus causing all of those tests to fail.
>>  
>>  I worked around this problem and managed to get the build to finish by
>>  moving gcc.exe to gcc.real and installing this as gcc:
>>  
>>    #!/bin/sh
>>  
>>    gcc.real "$@" 2>/tmp/err.$$
>>    STATUS=$?
>>    grep -v "file path prefix \`.*' never used" /tmp/err.$$ 1>&2
>>    rm -f /tmp/err.$$
>>    exit $STATUS
>>  
>>  Is this something y'all expect to fail?  If so, is it documented
>>  anywhere?  If not, can it be fixed?
>
>If anyone answered this part of my message, I didn't see it; did some
>send an answer to the mailing list without CC'ing it to me?
>
>In any case, this is still broken in the 10/18 snapshot.
>
>Can someone from the Cygwin team either acknowledge that there is a
>problem here or explain why it isn't one? :-)

There is a problem.  Someone changed the top level configure a while ago
and broke things.  Newer versions of gcc won't output the warning that you're
seeing so configure succeeds.

There hasn't been much activity lately in coming up with a fix that is
acceptable for older compilers so, if you have a generic fix, it would
be appreciated.  Otherwise your work around probably makes sense.

cgf

--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe AT sourceware DOT cygnus DOT com