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
X-Authentication-Warning:	tweedle.cabbey.net: cabbey owned process doing -bs
Date:	Tue, 30 Jan 2001 16:25:54 -0600 (CST)
From:	Christopher Abbey <cabbey AT bresnanlink DOT net>
X-Sender:	cabbey AT tweedle DOT cabbey DOT net
To:	cygwin AT sources DOT redhat DOT com
Subject:	Re: -mno-cygwin and c++ stl... progress?
In-Reply-To:	<Pine.HPP.3.96.1010130143255.2988H-100000@hp2.xraylith.wisc.edu>
Message-ID:	<Pine.LNX.4.21.0101301615470.3671-100000@tweedle.cabbey.net>
MIME-Version:	1.0

Today, Mumit Khan wrote:
> I'm not I understand what you're referring to. Contrary to some
> opinion out there, -mno-cygwin *in general* does not work for C++
> code unless you have the appropriate target libraries from Mingw 
> distribution. The reason is twofold -- (1) libgcc.a built for Cygwin 
> may or may reference __impure_ptr from Cygwin DLL even if you don't
> *think* you're using exceptions (long story, the keyword here is 
> inhibit_libc), and (2) the C++ runtime library for Cygwin and Mingw
> are *not* compatible.

wow. glad I asked... :)  #1 was news to me, #2 was the part I remember
talk of working around. I see you and cgf are exchanging ideas... I'll
test out the gcc he posted tonight when I get home.

> Please elaborate on the issues so that I can remedy these.

we're still poking around at makefiles and what not to figure out
just what's going wrong... but the highlevel sumary is you HAVE
to build in the source directory... seperating the trees causes
mingw's g++ to choke on reading some files... example below. If
we don't nail it down to something in our make files then one of
us will I'm sure submit to to the mingw list (if Mo doesn't I will)

/jikes/cvs/msvcbld/src $ make
/usr/local/mingw/bin/g++ -DHAVE_CONFIG_H -I. -I../../jikes/src -I.     -g
-O2 -c -o ast.o ../../jikes/src/ast.cpp
In file included from ..\..\jikes\src\platform.h:205,
                 from ..\..\jikes\src\ast.h:14,
                 from ../../jikes/src/ast.cpp:11:
..\..\jikes\src\tuple.h: In method `void Tuple<T>::AllocateMoreSpace()':
..\..\jikes\src\tuple.h:55: unknown escape sequence `\.'
..\..\jikes\src\tuple.h:55: unknown escape sequence `\j'
..\..\jikes\src\tuple.h:55: unknown escape sequence `\s'
In file included from ../../jikes/src/ast.cpp:11:
..\..\jikes\src\ast.h: In method `void StoragePool::AllocateMoreSpace()':
..\..\jikes\src\ast.h:4124: unknown escape sequence `\.'
..\..\jikes\src\ast.h:4124: unknown escape sequence `\j'
..\..\jikes\src\ast.h:4124: unknown escape sequence `\s'
make: *** [ast.o] Error 1

The same tree builds fine under cygwin or linux (it's a network mount)
The exact "escape sequence" has changed a couple times, but the locations
are consistent. (tuple.h:55 happens to be a comment block.)

-- 
now the forces of openness have a powerful and
  unexpected new ally - http://ibm.com/linux



--
Want to unsubscribe from this list?
Check out: http://cygwin.com/ml/#unsubscribe-simple

- Raw text -