Mailing-List: contact cygwin-help AT sourceware DOT cygnus DOT com; run by ezmlm List-Subscribe: List-Archive: List-Post: List-Help: , Sender: cygwin-owner AT sources DOT redhat DOT com Delivered-To: mailing list cygwin AT sources DOT redhat DOT com Date: Sat, 17 Mar 2001 14:45:18 +0100 (MET) From: Hubert Garavel Message-Id: <200103171345.OAA20679@ubac.inrialpes.fr> Reply-To: Hubert Garavel X-Mailer: Mail User's Shell (7.2.5 10/14/92) To: cygwin AT sources DOT redhat DOT com Subject: fstat problem: linking with crt or msvc? Cc: Hubert DOT Garavel AT inrialpes DOT fr, earnie_boyd AT yahoo DOT com I face a backward compatibility problem with fstat() due to a recent change in gcc -mno-cygwin (the new GCC links against MVSC instead of CRT). There have been already a few posts on this issue, but my question is different. I have a binary library (".a" file) which calls fstat(). This library was obtained by cross-compilation using an older GCC that links against CRTDLL (gcc version 2.95.2 19991024 (release-2). This library worked well on previous versions of Cygwin, but fails on the recent ones (lasting from 2001). It seems that, when linked to MSVC, fstat() uses different data structures and returns different results than when linked to CRT. If possible, I would keep the .a library unchanged. Which options should be given to the new gcc (in addition -mno-cygwin) so that it continues to link against CRTDLL as before and uses the appropriate include files for this? This is really a backard compatibility model. Thanks in advance Hubert PS: For my curiosity, what are the reasons for switching from CRT to MSVC ? -- ',',',',',',',',',',',',' Hubert GARAVEL | INRIA Rhone-Alpes / VASY ',',',',',',',',',',',',' hubert DOT garavel AT inria DOT fr | 655, avenue de l'Europe ',',',',',',',',',',',',' tel: +(33) 4 76 61 52 24 | 38330 Montbonnot St Martin ',',',',',',',',',',',',' fax: +(33) 4 76 61 52 52 | France ',',',',',',',',',',',',' http://www.inrialpes.fr/vasy -- Want to unsubscribe from this list? Check out: http://cygwin.com/ml/#unsubscribe-simple