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: <3BEF54C5.1060102@ece.gatech.edu>
Date: Sun, 11 Nov 2001 23:49:09 -0500
From: Charles Wilson <cwilson AT ece DOT gatech DOT edu>
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:0.9.4) Gecko/20011019 Netscape6/6.2
X-Accept-Language: en-us
MIME-Version: 1.0
To: Jason Tishler <jason AT tishler DOT net>
CC: cygwin-apps AT cygwin DOT com
Subject: Re: Python _cursesmodule build problem (was Re: [RFD] ncurses update)
References: <20011111230534 DOT C1876 AT dothill DOT com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

Jason Tishler wrote:

> Chuck,
> 
> On Sat, Sep 15, 2001 at 03:03:16PM -0400, Charles Wilson wrote:
> 
>>#1. Now use the auto-import stuff instead of __declspec(dllXXport) 
>>modifiers.  No more need for *compile-time* flags at all; only linktime 
>>flags: -static or -Wl,--enable-auto-import.  However, ncurses does some 
>>of the "bad" things (the ones that today's binutils warns about, but the 
>>previous binutils created buggy code for).  The fix for this requires 
>>changes in the ncurses library code (specifically, configure using 
>>BROKEN_LINKER -- which changes some of the exported DATA into functions. 
>>  If you recompile apps, you don't need to worry about this though; it's 
>>
>                                ^^^^
> 
>>transparently handled in ncurses_cfg.h/curses.h).
>>
> 
> I submitted the attached patch in order for Python (specifically the
> _cursesmodule module) to build cleanly again under the latest binutils.
> The above seems to imply that I shouldn't need this patch.  Did I
> misinterpret the above?


No, you didn't misinterpret.  I made a mistake.  However, the #define 
BROKEN_LINKER shouldn't go into client code, it *should* be in ncurses.h 
(or ncurses_cfg.h, etc).  In fact, I thought it WAS.  But on closer 
inspection, it looks like the appropriate definition only appears in 
private headers that aren't shipped with the compiled ncurses package.

I need to fix this on the ncurses side.

--Chuck