Date: Thu, 18 Jan 1996 01:00:40 -0800 From: "Jordan K. Hubbard" <jkh@time.cdrom.com> To: "Julian H. Stacey" <jhs@freebsd.org> Cc: Poul-Henning Kamp <phk@critter.tfs.com>, hackers@freebsd.org, Ollivier Robert <roberto@keltia.freenix.fr>, joerg@freebsd.org Subject: Re: Building a "custom" release of 2.1.0 Message-ID: <23332.821955640@time.cdrom.com> In-Reply-To: Your message of "Wed, 17 Jan 1996 14:52:47 %2B0100." <199601171352.OAA02448@vector.jhs.local>
next in thread | previous in thread | raw e-mail | index | archive | help
> If FreeBSD were doing a 2 stage {make all , install, chroot, remake }
> I wouldn't be querying the time taken, it's just the mandatory
> inbuilt CVS that's not obvious.
Heh? It's *extemely* obvious, Julian! You're just not looking at the
whole process. Think, man. How else do you generate a source tree to
build from in the first place that you *know* is pristine and, what's
more, conforms to a given release tag in the tree? We have the
capability to generate a release at any revision level, you know, not
just the current "head" of the source tree. When I start doing my 2.1
snaps, I'll be generating a source tree from the 2.1-STABLE branch.
If I wanted to roll another 2.1-RELEASE (for whatever reason) I could
do that too. When I roll the 2.2-SNAPs, they'll be from yet a
different revision tag (the HEAD).
OK, I could decouple it and make this a manual step, but then so could
anyone else who really wanted to do a release without CVS. It
certainly wouldn't help ME or any of the other release engineers to do
so, so we don't as a matter of general policy. If you or anyone else
wants to decouple CVS from the release process, it's like a 2 line
change (and if you don't understand /usr/src/release/Makefile
backwards and forwards anyway, believe me - you have no business even
trying to make a release yet! More understanding than knowing to type
`make' is required).
Jordan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?23332.821955640>