Date: Tue, 9 Jun 1998 11:24:45 -0400 (EDT) From: woods@zeus.leitch.com (Greg A. Woods) To: freebsd-stable@FreeBSD.ORG Subject: Re: Matt Behrens: Re: kernel compile problem Message-ID: <199806091524.LAA10150@brain.zeus.leitch.com> In-Reply-To: Richard Wackerbarth's message of "Mon, June 8, 1998 17:18:21 -0500" regarding "Re: Matt Behrens: Re: kernel compile problem" id <l03130302b1a211e9952c@[208.2.87.10]> References: <l03130300b19e494c0321@[208.2.87.10]> <199806052348.JAA10962@gsms01.alcatel.com.au> <199806081512.LAA16030@brain.zeus.leitch.com> <l03130302b1a211e9952c@[208.2.87.10]>
next in thread | previous in thread | raw e-mail | index | archive | help
[ On Mon, June 8, 1998 at 17:18:21 (-0500), Richard Wackerbarth wrote: ] > Subject: Re: Matt Behrens: Re: kernel compile problem > > The fundamental problem is that there are not enough "branches". As > a result, the RELENG_2_2 branch is having to serve as both the > development branch for 2.2.7 and the patch branch for 2.2.6. > For the former purpose, ASAP is a good strategy. For the latter, > any change to the tools destroys the ability of a user to > apply patches to his kernel. I agree that more branches might be better, but more branches require more management, and I'm not sure that would be better for a mostly volunteer project. I've found that when you try to do release branch management on top of other ongoing development branches you pretty well have to have one or more people dedicated to repository management and maintenance, and it really is a *full* time job. The "freeze development now and begin to prepare release directly on the branch" mechanism works well enough, esp. for a -stable branch where the *real* development does happen on a separate branch. > I don't think that the "kernel only" source users would consider > my argument at all bogus. For them, the 2.2 branch is BROKEN. They > cannot apply the latest security updates which affect the kernel > until 2.2.7 is released. It's only broken until it's fixed, and the sooner it's fixed, the better, as then more people will be able to test them. The obvious fix has been discussed, and it seems as if it will be implemented. It has proven extremely effective in NetBSD. (I.e. to include src/usr.sbin/config and its contents in the kernel-only source distrbution collections available through ftp, sup, CVSup, CTM, etc.) BTW, anyone seriously trying to keep right up to the minute with security fixes and such would be better off with user-land only than with kernel-only. People with so few resources should let other people manage the source tree and they should stick to releases only. Working with a live CVS repository (or even a daily copy of it) is bound to produce unpredictable results. Anyone crying wolf at the slightest perturbance will go hoarse and loose their voice long before they become too annoying. ;-) -- Greg A. Woods +1 416 443-1734 VE3TCP <gwoods@acm.org> <robohack!woods> Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com> To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199806091524.LAA10150>