From owner-freebsd-current@FreeBSD.ORG Mon Jan 7 07:49:23 2008 Return-Path: Delivered-To: freebsd-current@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id D867C16A418; Mon, 7 Jan 2008 07:49:23 +0000 (UTC) (envelope-from danny@cs.huji.ac.il) Received: from cs1.cs.huji.ac.il (cs1.cs.huji.ac.il [132.65.16.10]) by mx1.freebsd.org (Postfix) with ESMTP id 9775713C467; Mon, 7 Jan 2008 07:49:23 +0000 (UTC) (envelope-from danny@cs.huji.ac.il) Received: from pampa.cs.huji.ac.il ([132.65.80.32]) by cs1.cs.huji.ac.il with esmtp id 1JBmjZ-00047M-05; Mon, 07 Jan 2008 09:49:21 +0200 X-Mailer: exmh version 2.7.2 01/07/2005 with nmh-1.2 To: Andrew Reilly In-reply-to: Your message of Mon, 7 Jan 2008 07:42:09 +1100 . Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: quoted-printable Date: Mon, 07 Jan 2008 09:49:20 +0200 From: Danny Braniss Message-ID: Cc: rgrav , Tim Kientzle , freebsd-current@freebsd.org, Peter Schuller , Jason Evans Subject: Re: ELF dynamic loader name [was: sbrk(2) broken] X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 07 Jan 2008 07:49:23 -0000 > On Sun, 06 Jan 2008 09:56:32 +0200 > Danny Braniss wrote: >=20 > > what Apple has is one file, that will run the appropiate binary if ru= n > > on an i386 or a ppc, not 2 different files - universal binary - not r= osetta. >=20 > Sure, but that's got a bunch of different driving factors. I > don't know, for example, whether you can build a four-way > executable (ia32, x86_64, ppc, ppc64). Well, you probably can, > but I'd be a bit surprised if anyone has. FreeBSD supports even > more architectures: it just doesn't scale. The best bet for > something that has to run everywhere is probably LLVM or TNEF. >=20 > The advantage that Unix has over MacOS is that we aren't trying > to squeeze everything into single =22application=22 directories. So > it's reasonable to have =22share=22, and select executables on the > basis of PATH. That's how it has worked before. Most sites > don't have more than two or three different architectures to > support, anyway. >=20 This argument has sides/issues, one is the 'distribution', and here I agr= ee that one universal-fit-all is not the way to go. I'm concerned in trying to solve a problem we are facing here, were=20 students/researchers write code, and soon will be hit by incompatible platforms. > If we do get much further with multi-architecture bin and lib, > and people actively use these on diskless setups or > multi-architecture hosts (amd64/ia32, or other 64/32 bit > combinations being the most common) then perhaps it would be nice > to have a share/bin where platform-independent scripts (shell, > perl, python) as well as dynamic-translated binaries (JVM, LLVM, > etc) can live? >=20 > Cheers, >=20 > --=20 > Andrew danny