From owner-freebsd-afs Fri Apr 20 11:24:50 2001 Delivered-To: freebsd-afs@freebsd.org Received: from zaphod.ece.cmu.edu (ZAPHOD.ECE.CMU.EDU [128.2.136.35]) by hub.freebsd.org (Postfix) with ESMTP id 3CD5837B422 for ; Fri, 20 Apr 2001 11:24:45 -0700 (PDT) (envelope-from tardis@watson.org) Received: (from tardis@localhost) by zaphod.ece.cmu.edu (8.11.0/8.8.8) id f3KIOgG29883; Fri, 20 Apr 2001 14:24:42 -0400 (EDT) X-Authentication-Warning: zaphod.ece.cmu.edu: tardis set sender to tardis@watson.org using -f To: freebsd-afs@FreeBSD.org, port-freebsd@openafs.org Subject: Build fix From: Tom Maher Date: 20 Apr 2001 14:24:41 -0400 Message-ID: Lines: 74 User-Agent: Gnus/5.0807 (Gnus v5.8.7) XEmacs/21.1 (Canyonlands) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: owner-freebsd-afs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org About a day after initial big patch was comitted, another change was comitted to pinstall/install.c, which is used only for the AFS build system, AFAIK. This had the sad effect of breaking the build on FreeBSD, as install started stripping everything in sight, including text files and library symbol tables. I've submitted the patch to openafs-gatekeepers, and you can find it at http://www.watson.org/~tardis/openafs/oafs-fbsd-2001-04-18.patch Much smaller this time, only 1.5 KB. Due to popular request, here are some simple build instructions. You can either slurp down CVS, or just grab one of the daily CVS snapshot tarballs, conveniently located right off of http://www.openafs.org/fset.shtml/release/snapindex.html For grabbing it via anon CVS, set CVSROOT to :pserver:anonymous@www-openafs.central.org:/cvs and then `cvs checkout openafs`. One of the openafs elders (Derrick Brashear, I think) said after 1.0.4, they'll be forking the versions (1.0.x as stable, 1.1.x as devel), and then doing the standard stable-as-even, devel-as-odd thing. FreeBSD being a new port, the work will mostly all be in the devel version. For building on local disk, the routine looks something like $ cd /path/to/openafs # you should be at same level as "src" and "doc" $ patch -p0 < /path/to/oafs-2001-04-18.patch # patch may be already comitted when you read this # check http://www.watson.org/~tardis/openafs/ for "LATEST_IS_" $ mkdir i386_fbsd_42 $ ln -s src/Makefile $ ln -s i386_fbsd_42 @sys $ ln -s @sys/obj $ ln -s @sys/dest $ gmake SYS_NAME=i386_fbsd_42 links $ gmake SYS_NAME=i386_fbsd_42 You'll want to have GNU bison around, and you'll want to use GNU make, not system make. Experienced AFS users will note the @sys hackery. If you want to build in AFS with arla, you'll need to have your sysname set to "i386_fbsd_42" (`fs sysname` gets your current sysname, `fs sysname i386_fbsd_42`, as root, will set it to i386_fbsd_42). Once 4.3-RELEASE hits my favorite mirror, I'll fix the config files so both i386_fbsd_42 and i386_fbsd_43 will work. The AFS build system is currently extremely crufty, makes a number of assumptions, and hacks around the fact that "washtool", a vital utility used during the build, isn't there because IBM didn't release the source. This makes it very hard to build individual parts of the tree without building the entire tree. Parts of the tree which will be of particular interest will be src/afs/FBSD, src/afs/VNOPS, and src/rx/FBSD, which is where a bunch of preliminary kernel code lives. By "preliminary", I mean "copied out of the DUX directories". The configuration headers for FreeBSD are in src/config/param.i386_fbsd_42*.h. Since I've been naughty and given almost no consideration to anything but FreeBSD/x86, people interested in FreeBSD/Alpha, and who have alphas running FreeBSD, could be useful. Feedback is mega-welcome, and direct patches go to openafs-gatekeepers@openafs.org (a private list for the comitters, of which I am not one). I'd also really appreciate it if people cc'd this list. Also, please feel free to bug me if you can make the code work but don't feel comfortable with diff/patch. -- Tom Maher To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-afs" in the body of the message