Mailing-List: contact cygwin-apps-help AT sourceware DOT cygnus DOT com; run by ezmlm
Sender: cygwin-apps-owner AT sourceware DOT cygnus DOT com
List-Subscribe: <mailto:cygwin-apps-subscribe AT sources DOT redhat DOT com>
List-Archive: <http://sources.redhat.com/ml/cygwin-apps/>
List-Post: <mailto:cygwin-apps AT sources DOT redhat DOT com>
List-Help: <mailto:cygwin-apps-help AT sources DOT redhat DOT com>, <http://sources.redhat.com/lists.html#faqs>
Delivered-To: mailing list cygwin-apps AT sources DOT redhat DOT com
Message-ID: <20011107211325.50364.qmail@web14504.mail.yahoo.com>
Date: Thu, 8 Nov 2001 08:13:25 +1100 (EST)
From: =?iso-8859-1?q?Danny=20Smith?= <danny_r_smith_2001 AT yahoo DOT co DOT nz>
Subject: Re[2]: Want to be a gcc/binutils maintainer?
To: cwilson AT ece DOT gatech DOT edu
Cc: egor duda <cygwin-apps AT cygwin DOT com>
In-Reply-To: <20011107204726.16501.qmail@web14505.mail.yahoo.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit


> > 
> > Question, Danny:  with your changes, does auto-import continue to work?
> Yes, in all my tests.
> 
> > (with *non* fastcall symbols only, for now...auto-import-for-fastcall
> > probably needs to be added later, but for now...)
> > 
> 
> fastcall symbols are functions, never data, so no need to auto-import. 
> However, all the _nm_ symbols seem to be okay.
> 
> 
> > Download this:
> > 
> > 
> >
>
http://www.neuro.gatech.edu/users/cwilson/cygutils/V1.1/dll-stuff/dll-helpers-0.2.9.tar.bz2
> > 
> > and run the various tests.
> > 
> > 
> I get broken link messages with that.  I'll try later. FWIW,
> dllhelpers-0.2.4-0.2.8 all work before and after changing functions to
> __fastcall (and with __fastcall C++ methods), and mixing __stdcall,
> __cdecl, and __fastcall in same code.    I don't know how to mark fortran
> names as fastcall.
> 
> --Chuck 
> 
> 

One "auto-feature" that is not yet enabled is extension of
--enable-stdcall-fixup to fastcall.  Fixup of fastcall names to cdecl names
is relatively simple.  Fixup of cdecl names to fastcall is more expensive
since it means looking up '@foo@' as well as 'foo@' whenever there is an
undefined foo. Also should we try to fixup up undefined fastcall @foo AT n to
defined stdcall foo AT n (and vise versa)?  I've left all that out for now so
we get hard link errors rather than fixups if (when) things go wrong
somewhere else.

Danny

http://briefcase.yahoo.com.au - Yahoo! Briefcase
- Manage your files online.