Mailing-List: contact cygwin-developers-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-developers-subscribe AT cygwin DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin-developers/>
List-Post: <mailto:cygwin-developers AT cygwin DOT com>
List-Help: <mailto:cygwin-developers-help AT cygwin DOT com>, <http://sources.redhat.com/ml/#faqs>
Sender: cygwin-developers-owner AT cygwin DOT com
Delivered-To: mailing list cygwin-developers AT cygwin DOT com
Message-ID: <002301c24e2f$a53971f0$6132bc3e@BABEL>
From: "Conrad Scott" <Conrad DOT Scott AT dsl DOT pipex DOT com>
To: <cygwin-developers AT cygwin DOT com>
References: <000201c24e27$ffa0bb10$6132bc3e AT BABEL> <20020828003122 DOT GL16631 AT redhat DOT com>
Subject: Re: A quick note on <newlib.h>
Date: Wed, 28 Aug 2002 02:10:05 +0100
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000

"Christopher Faylor" <cgf AT redhat DOT com> wrote:
> This is one of the reasons that I'm getting sick of our
dependency on
> newlib.  I've asked that cygwin be taken into account when
making
> changes like this but, the last I heard, the newlib guys were
stalled
> trying to accommodate my request since the two year old gcc
cross
> compiler that they insist on using is no longer able to build
cygwin.

Ick.  Sounds like an easy way to stall lots of things . . .

> It's easy enough to add another include path to cygwin but I'm
not sure
> that I want to do that.  I think, instead, I'm going to start
thinking
> about how we can eliminate our dependency on newlib.

It would be good to disentangle the build process so that a cygwin
build didn't have to add include paths across to the newlib source
tree.  This sounds a bit like the "why does cygwin use -nostdinc"
discussion that flickered for a while last month when the <new.h>
issue in cygserver showed up.  It wouldn't seem impossible to rely
on installed copies of the newlib and liberty libraries and their
header files.  The cygwin build wouldn't then need to reach
outside the winsup source tree, and if that were included first it
would seem to obviate the need for the -nostdinc flag.  I recall
you had worries about not using -nostdinc that currently I can't
recall, so this may be a no-hoper. It's also just an off the top
of the head idea tho' and I've not looked into this, except for a
little while earlier on trying to patch the gcc specs file for my
local setup and to understand the <newlib.h> header file issue.

> In the meantime, I've checked in an empty "newlib.h" file into
the
> winsup/cygwin.  That should allow things to build, at least.

Thanks.

// Conrad