Mailing-List: contact cygwin-apps-help AT cygwin DOT com; run by ezmlm
Sender: cygwin-apps-owner AT cygwin DOT com
List-Subscribe: <mailto:cygwin-apps-subscribe AT cygwin DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin-apps/>
List-Post: <mailto:cygwin-apps AT cygwin DOT com>
List-Help: <mailto:cygwin-apps-help AT cygwin DOT com>, <http://sources.redhat.com/lists.html#faqs>
Delivered-To: mailing list cygwin-apps AT cygwin DOT com
Date: Mon, 21 Jan 2002 22:42:31 -0500
From: Christopher Faylor <cgf AT redhat DOT com>
To: cygwin-apps AT cygwin DOT com
Subject: Re: gcc v3 issue -- hacky solution
Message-ID: <20020122034231.GA23893@redhat.com>
Reply-To: cygwin-apps AT cygwin DOT com
Mail-Followup-To: cygwin-apps AT cygwin DOT com
References: <20020122032839 DOT GA23746 AT redhat DOT com> <08d401c1a2f4$d2d66370$0200a8c0 AT lifelesswks>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <08d401c1a2f4$d2d66370$0200a8c0@lifelesswks>
User-Agent: Mutt/1.3.23.1i

On Tue, Jan 22, 2002 at 02:28:12PM +1100, Robert Collins wrote:
>
>===
>----- Original Message -----
>From: "Christopher Faylor" <cgf AT redhat DOT com>
>To: <cygwin-apps AT cygwin DOT com>
>Sent: Tuesday, January 22, 2002 2:28 PM
>Subject: gcc v3 issue -- hacky solution
>
>
>>If we can generate new operators for all of the builtin types that are
>>currently being used in setup.exe, then, if we compile everything with
>>-fno-exceptions (which I've just added to Makefile.in) we won't have
>>to worry about having a mingw version of libstdc++.a around.
>>
>>It's a hack, I know, but the alternative of somehow building a mingw
>>version of the stdc library is not too attractive.
>>
>>So, if anyone was looking for something, oh I don't know, repetitive to
>>do, that would be a good project, assuming that Robert agrees.
>>
>>(Unfortunately, my internal Robert simulator gives this about a 30%
>>chance of an agreement)
>
>This external Robert simulator can confirm that, and drop it to about
>10% :}.
>
>I want to enable exceptions, to cleanup more of the code.  I wouldn't
>be stressed about haveing all the operators (surely there's a library
>around already),

There's a library -- it's libstdc++.a.

>but the no-exceptions clause won't fly without some convincing.

How about, you can't do that unless you come up with some way of
ensuring that there is a mingw version of libstdc++ available?

The code in the sources.redhat.com:/cvs/src repository is supposed to be
self-hosting.  You should be able to build it with nothing but a native
C compiler/linker.  We can't just say "You have to have the right
library available."

That seems like a pretty steep hill to me.

Or, maybe we move the cinstall directory out of the winsup hierarchy
and into cygwin-apps.  Then we can make our own rules.

Please don't get me wrong.  It would actually make my life much easier
if we could somehow create a mingw version of libstdc++.a (I'm getting
carpal tunnel from typing that so much...).  The reason I'm bringing
this up now is that I'm getting internal "This isn't building" email
about some of the things in winsup/utils.  This wouldn't be an issue
at all if we had a mingw version of libstdc+...  Ow!

DJ is checking on possibly building... that library... as a multilib.
I don't know how successful this will be however.

cgf