From owner-freebsd-current Wed Mar 6 18:31:53 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.3/8.7.3) id SAA03178 for current-outgoing; Wed, 6 Mar 1996 18:31:53 -0800 (PST) Received: from dyson.iquest.net (dyson.iquest.net [198.70.144.127]) by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id SAA03171 Wed, 6 Mar 1996 18:31:47 -0800 (PST) Received: (from root@localhost) by dyson.iquest.net (8.6.11/8.6.9) id VAA02388; Wed, 6 Mar 1996 21:27:24 GMT From: "John S. Dyson" Message-Id: <199603062127.VAA02388@dyson.iquest.net> Subject: Re: Paul's rules of order for committers (Re: -current submitting policys) To: Daniel.M.Obrien@att.com Date: Wed, 6 Mar 1996 21:27:24 +0000 () Cc: ejc@naserver1.cb.att.com, pst@shockwave.com, freebsd-current@FreeBSD.org, joerg_wunsch@uriah.heep.sax.de, committers@FreeBSD.org, terry@lambert.org, jkh@time.cdrom.com In-Reply-To: <9603062141.AA10817@cbsky.cb.att.com> from "Daniel.M.Obrien@att.com" at Mar 6, 96 04:41:51 pm X-Mailer: ELM [version 2.4 PL24 ME8] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-current@FreeBSD.org Precedence: bulk > > I guess it is good that the ``fix'' is turned around quickly, but what bothers > me is that I'm left with the feeling that the source submission process > (source tree) is being used as a place to park changes rather than parking > them locally (on developer's system), and compiling & soaking them on the > developer's system. (Thus bending Paul's unwritten rules above.) > Unfortunately, the problem that we have is that each developer has only one, two, and perhaps fortunate ones have three systems to develop on. The VM changes that I added had been run on my system for about 2wks, and sent them to another person for them to use for about 1wk. Problems leak through, and I try to fix the problems that I create very quickly. I have some (evil) soak tests that I run, much worse than the SVR4 tests that I used to run at work (SVR4 would practically stop or break under the load that I put my system under.) -current is not the latest working code, but the latest attempted working code. Those that run -current in production need to track it very carefully and grab stable copies. We don't have a good, formal policy for backing changes out, and I am willing to back the VM changes out in a few days if the system continues to break. In fact, it might be useful to pull a few back one-by-one until the problems go away. AFAIK, -stable is the only "released" code. John