Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
Message-ID: <41084193.60206@nycap.rr.com>
Date: Wed, 28 Jul 2004 20:15:15 -0400
From: Mathieu Malaterre <mmalater AT nycap DOT rr DOT com>
Reply-To: mathieu AT malaterre DOT com
User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.6) Gecko/20040113
MIME-Version: 1.0
To: Charles Wilson <cygwin AT cwilson DOT fastmail DOT fm>
CC: cygwin AT cygwin DOT com
Subject: Re: INT32: jmorecfg.h
References: <4105AB1C DOT 8060006 AT nycap DOT rr DOT com> <NUTMEG6PQsqRje8toOH00000aa3 AT NUTMEG DOT CAM DOT ARTIMI DOT COM> <20040727150250 DOT GC16364 AT trixie DOT casa DOT cgf DOT cx> <41073BD1 DOT 3060707 AT cwilson DOT fastmail DOT fm>
In-Reply-To: <41073BD1.3060707@cwilson.fastmail.fm>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
X-Virus-Scanned: Symantec AntiVirus Scan Engine
Note-from-DJ: This may be spam

Charles Wilson wrote:

> Christopher Faylor wrote:
> 
>> On Tue, Jul 27, 2004 at 10:01:08AM +0100, Dave Korn wrote:
>>
>>> I remember hitting this problem with a cygwin build of libjpeg once.
>>> The correct answer is probably that jmorecfg.h shouldn't arbitrarily
>>> define INT32 globally without first performing an autoconf test to see
>>> if there's something in the system includes that already provides it,
>>> isn't it?
>>
>>
>>
>> Yes.  Just adding more and more ifdef guards around the INT32 definition
>> is not the way to go.
> 
> 
> Unfortunately, the ijg group does not distribute the autoconf source for 
> libjpeg's configure script.  It's on my todo list to really truly 
> autoconfiscate libjpeg (and libtiff) -- which will make things like this 
> easier.  But it's not a very high priority.
> 
> Nor am I all that worked up about this INT32 issue.  jmorecfg.h is 
> present in all builds of libjpeg; I don't understand why INT32 is 
> causing a problem *only* on cygwin.  It seems that you should be seeing 
> the problem on all platforms, or none.  Since your package is showing 
> this behavior only on cygwin, I suspect something wacky in (a) your 
> cygwin environment, or (b) the procedure/configury you're using to build 
> on cygwin (e.g. mixing native and cygwin headers).
> 
> Given that, I'm not going to rush out and fix this right away, because I 
> only have bad choices: (1) add yet another #ifdef hack, or (2) spend a 
> week teasing out and reverse-engineering the autoconfigury source files...

Alright,

	Here is what we did to work around this problem instead of directly 
include jpeglib.h we do:

/* Work-around for a conflict between windows.h and jpeglib.h in cygwin.
If ADDRESS_TAG_BIT is defined then BaseTsd.h has been included and
INT32 has been defined with a typedef, so we must define XMD_H to
prevent the jpeg header from defining it again.  */

# if defined(__CYGWIN__) && defined(ADDRESS_TAG_BIT) && !defined(XMD_H)
#  define XMD_H
#  define VTK_JPEG_XMD_H
# endif
# include <jpeglib.h> 	 # include <jpeglib.h>
# if defined(VTK_JPEG_XMD_H)
#  undef VTK_JPEG_XMD_H
#  undef XMD_H
# endif

http://vtk.org/cgi-bin/viewcvs.cgi/Utilities/vtk_jpeg.h?r1=1.1&r2=1.2&root=VTK

The whole problem was that one can't do:

#include <windows.h>
#include <jpeglib.h>

And according to Christopher, this is "not designed to work well.". Thus 
I don't believe the mistake is in VTK. My guess is if we don't see in 
other plateform is because they -luckily- don't use any typedef INT32.

HTH
Mathieu

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/