From owner-freebsd-arch Thu Jan 31 16:53:12 2002 Delivered-To: freebsd-arch@freebsd.org Received: from avocet.prod.itd.earthlink.net (avocet.mail.pas.earthlink.net [207.217.120.50]) by hub.freebsd.org (Postfix) with ESMTP id C00E137B402 for ; Thu, 31 Jan 2002 16:53:09 -0800 (PST) Received: from pool0167.cvx40-bradley.dialup.earthlink.net ([216.244.42.167] helo=mindspring.com) by avocet.prod.itd.earthlink.net with esmtp (Exim 3.33 #1) id 16WRwl-0006Tw-00; Thu, 31 Jan 2002 16:52:56 -0800 Message-ID: <3C59E6E2.BFC3836A@mindspring.com> Date: Thu, 31 Jan 2002 16:52:50 -0800 From: Terry Lambert X-Mailer: Mozilla 4.7 [en]C-CCK-MCD {Sony} (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Julian Elischer Cc: Sheldon Hearn , Marco Molteni , arch@FreeBSD.org Subject: Re: Adding support for a global src tree serial number References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-arch@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Julian Elischer wrote: > Basically you are talking about an extension to the commit logging > process. > > it would be good if the file also listed all files changed in that commit. The problem is a little larger. o What if I'm on another branch? o What if my source code was checked out last Tuesday, and I've continued to CVSup? o What if I told CVS to check out by date, to ensure I got the code from last Tuesday? The last one is the kicker, since it's what people are going to use to reproduce problems, since really, what they want is a "check out by serial number" so they can see the problems someone is having with a given serial number, and fix them. In other words, the serial number can't just be HEAD-invariant per branch. Nate's approach of raw reading and writing the RCS file is the only workable one, I think, and there are still race windows for a partially eddited file obtained via CVSup. The only way around this is a two stage commit replacement of the file, and that leaves a sliver of a window where there is a "file.old" and a "file.new", without a "file" linked to one or the other, unless a hard link is allowed to replace a file, and CVSup will probably barf on it not being there, so it would have to be modified to back off and retry. This is basically the reason I dsuggested a daemon, in my previous posting. -- Terry To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message