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
Message-ID: <3B453DC3.4030201@ece.gatech.edu>
Date: Fri, 06 Jul 2001 00:25:39 -0400
From: "Charles S. Wilson" <cwilson AT ece DOT gatech DOT edu>
User-Agent: Mozilla/5.0 (Windows; U; WinNT4.0; en-US; rv:0.9.1) Gecko/20010607 Netscape6/6.1b1
X-Accept-Language: en-us
MIME-Version: 1.0
To: "jorgens AT coho DOT net" <jorgens AT coho DOT net>
CC: "Cygwin List (E-mail)" <cygwin AT cygwin DOT com>
Subject: Re: Cygwin 2.1 and Python modules
References: <01C1055C DOT 05CA03E0 DOT jorgens AT coho DOT net>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

That is *my* opinion, yes.  There are others who disagree.  Once the 
need to decorate (__declspec(dllexport)) both functions AND variables is 
no longer required -- because perhaps Paul S's changes are incorporated 
-- then I'd say SURE, ignore implibs and just link directly to the dll.

But not yet.

So help test Paul's patches.

--Chuck

Steve Jorgensen wrote:

> I see.  So for most of us, it's better to pretend this capability does not exist since it will only cause us headaches.
> 
> -----Original Message-----
> From:	Charles Wilson [SMTP:cwilson AT ece DOT gatech DOT edu]
> Sent:	Thursday, July 05, 2001 1:08 PM
> To:	jorgens AT coho DOT net
> Cc:	cygwin AT cygwin. Com (E-mail)
> Subject:	Re: Cygwin 2.1 and Python modules
> 
> Steve Jorgensen wrote:
> 
>>-----Original Message-----
>>From:   Jason Tishler [SMTP:Jason DOT Tishler AT dothill DOT com]
>>Sent:   Thursday, July 05, 2001 11:33 AM
>>To:     David Abrahams
>>Cc:     Cygwin
>>Subject:        Re: Cygwin 2.1 and Python modules
>>
>>Dave,
>>
>>...
>>It takes advantage of somewhat recent Cygwin gcc enhancements:
>>
>>    1. binding to msvcrt.dll instead of crtdll.dll
>>    2. ability to link directly to a DLL (instead of requiring an import
>>       library)
>>
>>I recently printed the section of the Cygwin User's Guide, and it appears
>>to show only the import-library method of linking to a DLL (if I understand
>>what I'm seeing).  I suppose it's important to keep it there since some
>>users may freeze their Cygwin installations for periods of time, but it
>>would be nice to have an example of the new way in the document as well.
>>
> 
> FYI, the "link directly to a DLL without using an import lib" is
> intended ONLY as an emergency fallback in those cases (MS DLL's) where
> implibs are not available.  This is based on the DJ's statements "back
> then" when he was implementing this functionality.
> 
> Also AFAIRC, "direct linking" doesn't even work if the DLL exports
> variables; only purely-functional interfaces are supported.  Now, Paul
> Sokolovsky has some patches to binutils that enable "non-decorated"
> import/export -- which conceivably can also enable "direct linking" with
> variable exports (since that's why Paul developed the patches). 
> However, Paul's changes have not yet been widely tested, and have not
> been accepted into the mainline binutils source code.  Check the mail
> archives for the patch, build a binutils (or use Paul's link to his
> prebuilt version) and help test it.
> 
> --Chuck
> 
> --
> Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
> Bug reporting:         http://cygwin.com/bugs.html
> Documentation:         http://cygwin.com/docs.html
> FAQ:                   http://cygwin.com/faq/
> 



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/