From: jeffdbREMOVETHIS AT goodnet DOT com (Mikey) Subject: Re: ld -r (relocatable link) 8 Apr 1998 18:59:46 -0700 Message-ID: <35312541.165056403.cygnus.cygwin32.developers@smtp.goodnet.com> References: <352A9273 DOT 8ACDDDB3 AT octanttech-nospam DOT com> Reply-To: jeffdbREMOVETHIS AT goodnet DOT com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit To: Simon Tuffs , cygwin32-developers AT cygnus DOT com I have been meaning to take a look at this, one Idea I had was to try a linker script for -r that moved .idata (and possibly .edata) into /DISCARD/ I believe that .idata is a large part of the reason -r dosen't work. On Tue, 07 Apr 1998 13:54:11 -0700, you wrote: >The GNU-WIN32 linker (ld) appears to be unable to perform relocatable >linking, and when faced with an object module built that way, it >crashes. I have seen others on this group comment that ld -r is not >supported, and I was wondering was there a technical reason, or >otherwise. I need this capability, and I will be happy to implement it >if I can, and if others can help me get started. I am currently working >around the problem by using the archiver to build archives rather than >relinkable object modules. > >For those interested, we need the relocatable object modules so we can >link together code which would otherwise declare duplicate functions >(with conflicting parameter types). By pre-linking the individual >modules before the final link we ensure that there are no symbol >conflicts at final link time. ===================================================== Linux a platform built by, and for users, standing on the firm legs of reliability, and speed. Microsoft Windows, a platform without a leg to stand on. (jeffdbREMOVETHIS AT netzone DOT com) delete REMOVETHIS from the above to reply Mikey