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: List-Archive: List-Post: List-Help: , Delivered-To: mailing list cygwin-apps AT sources DOT redhat DOT com Subject: Re: auto-import STATUS From: Robert Collins To: Charles Wilson Cc: Danny Smith , Paul Sokolovsky , cygwin-apps AT cygwin DOT com In-Reply-To: <3B6B6D15.1010804@ece.gatech.edu> References: <20010803233635 DOT 73332 DOT qmail AT web14506 DOT mail DOT yahoo DOT com> <996892555 DOT 8222 DOT 13 DOT camel AT lifelesswks> <3B6B6D15 DOT 1010804 AT ece DOT gatech DOT edu> Content-Type: text/plain Content-Transfer-Encoding: 7bit Message-Id: <996897738.11597.13.camel@lifelesswks> Mime-Version: 1.0 X-Mailer: Evolution/0.11 (Beta Release) Date: 04 Aug 2001 14:05:20 +1000 X-OriginalArrivalTime: 04 Aug 2001 03:51:13.0880 (UTC) FILETIME=[B4E81580:01C11C98] On 03 Aug 2001 23:33:41 -0400, Charles Wilson wrote: > Robert Collins wrote: > > > > It sure looks to me that you're wrong on this one. Pls give a code > reference or empirical evidence for this statement. I bow to thee. I haven't had my fingers deep in this for over a month. My empirical evidence was repeated attempts while working on libtool to prevent non-function .exe's. It sounds like I've got some wires crossed here, so I'll bow out at this point. What I thought happened was: the --export-all-symbols forced exporting _all_ symbols, not automatically filtered. This meant that we got _imp_ stuff from other .dlls (but as forwarded exports which would be ok) and that caused grief linking because libcygwin.a is appended to the list, meaning that linking an exe to a dll that depended on cygwin, resulted in things like __impure_ptr being resolved to the dll that depends on cygwin1.dll, not to cygwin1.dll itself. With libtool I found that after I patched ld to remove the core cygwin symbols I also had to _remove_ --export-all-symbols before it was corrected. Still, you have the source and more test results in front of you.. Rob > > My guess is there may still be a few bugs. :-) > > --Chuck > >