Subject: Re: $NetBSD$ Version Control Keyword
To: Curt Sampson <email@example.com>
From: Ken Hornstein <firstname.lastname@example.org>
Date: 07/22/1996 09:22:57
>Could someone please explain to this novice at version control why
>we use the keword $NetBSD$ instead of $Id$ in the NetBSD source
>files, how this is done with CVS, and what the ramifications are
>for users who want to use version control on NetBSD at home?
It's my understanding that this is done so people that want to incorporate
NetBSD sources into a local RCS/CVS tree won't change the version numbers
when local changes are made.
This has some advantages, as that means changes to check into RCS/CVS
locally won't change the version number, so that means diffs won't show
the change. Also, since the original version number is preserved, if/when
you submit changes to be made, they can more easily be merged if the original
revision number is known (not always necessary, but it is helpful at times).
I belive this is done by hacking RCS and CVS to recognize $NetBSD$ as
$Id$. Note that this is done by other people as well; the X Consortium
uses $XConsortium$ instead of $Id$
>I've got a personal patch or two patches to the NetBSD kernel and
>userland, and I'm getting a bit sick of sup wiping them out every
>night. I'd like to be able to hang on to them somehow. I'm thinking
>that putting the NetBSD revisions in the vendor branch of a CVS
>repository and then ocasionally doing an import from my sup copy
>might help here (though it still seems to mean that I've got to
>keep an entire separate copy of the source--*sigh*). Am I even
>heading in the right direction here?
Even worse, you'll need _three_ - one for CVS, one to SUP into, and a
checked-out copy to build NetBSD from.