From owner-freebsd-alpha Wed May 14 03:53:28 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id DAA10719 for alpha-outgoing; Wed, 14 May 1997 03:53:28 -0700 (PDT) Received: from spinner.DIALix.COM (spinner.dialix.com [192.203.228.67]) by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id DAA10712 for ; Wed, 14 May 1997 03:53:17 -0700 (PDT) Received: from spinner.DIALix.COM (localhost.dialix.com.au [127.0.0.1]) by spinner.DIALix.COM with ESMTP id SAA09926; Wed, 14 May 1997 18:51:16 +0800 (WST) Message-Id: <199705141051.SAA09926@spinner.DIALix.COM> X-Mailer: exmh version 2.0gamma 1/27/96 To: Curt Sampson cc: John Polstra , Warner Losh , freebsd-alpha@FreeBSD.ORG Subject: Re: Alpha questions.. In-reply-to: Your message of "Wed, 14 May 1997 03:14:23 MST." Date: Wed, 14 May 1997 18:51:15 +0800 From: Peter Wemm Sender: owner-freebsd-alpha@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk Curt Sampson wrote: > > I just wanted to know if it was complete > > (ie: fully functional with no missing parts, eg, 'ar' support). > > Ar support works, yes; we use it for static libraries. > > As I said, we have systems that run on ELF and nothing else, which > to me means that pretty much all support is there. If you have any > particular queries, I'm happy to answer them. Thanks.. I think I've found out what I was after now though.. > > Does it > > use elf-style shared libray versioning (ie: symlinks and internal names) > > or does it use a runtime ld.so.cache major.minor[.micro] a.out style > > system? > > It uses the elf-style system. However, there was some discussion > about how to deal with the semantics of using different versions > of the library, since the a.out system has different semantics from > the regular ELF system in terms of what it does if the desired > version is not available. Getting ELF to emulate this was looking > like a real PITA. I don't remember off-hand all the details of the > problem or the eventual result, however. I just had a rummage around, I can see the symlink stuff in bsd.lib.mk now, so I guess that answers my question.. It's using -soname with the libxx.so.major, with a symlink pointing to the current minor version. It looks like the pmax did a -soname of libxx.so.major.minor, but that wouldn't automatically find the highest minor version. Yes, it's a thorny problem.. > cjs Cheers, -Peter