Sender: bill AT taniwha DOT tssc DOT co DOT nz Message-ID: <35653DED.28DBED77@taniwha.tssc.co.nz> Date: Fri, 22 May 1998 20:57:18 +1200 From: Bill Currie MIME-Version: 1.0 To: DJ Delorie CC: andrewc AT rosemail DOT rose DOT hp DOT com, djgpp-workers AT delorie DOT com Subject: Re: underscores: a.out vs coff vs elf vs gcc vs as vs ld References: <199805212329 DOT TAA22346 AT delorie DOT com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Precedence: bulk DJ Delorie wrote: > If we do go through the hell of converting object formats again, we're > dumping the underscores. It will be slightly worse during the > conversion, but it will be better in the long run, especially if we > can be 100% compatible with Linux's ELF format. A little pain now saving a lot later? Makes sense to me, I only questioned the need to comply with ELF's underscore rule in my message in c.o.m.d. Anyway, we can always supply a program that converts assembly source from 'coff-go32' to 'elf32-i386' (ie goes through the source removing the leading undersore from global symbols) to help people (much like dtou and utod). Once I get ld working for `hppa1.1-stratus-sysv4' (got gcc working for both that target and 'i860-stratus-sysv4' (big endian i860, that was fun:/ )) I should know enough about ELF to get a loader going. I think I'll have gotten over my obsession with porting GNU stuff to the Stratus systems:). As to debuging support (symbols, not loading), is using libgdb a good or a bad idea? Seems to me it would be good from an implementation POV, but what about licencing? Hmm, back to the underscores issure...if we go to elf and drop the things, will that cause some of these gcc-as cross configuration problems to go away? Just for that ELF could be worth it. Bill -- Leave others their otherness.