Subject: Re: esoteric question: consistency of NetBSD CVS tree?
To: None <>
From: Bill Studenmund <>
List: current-users
Date: 10/16/2003 15:10:52
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Wed, Oct 15, 2003 at 04:31:14PM -0700, wrote:
> Please see:
> Now take a peek in the Attic.
> Notice there are two Makefiles--one in the normal sed/ and one sitting=20
> in the Attic. When a new version is added to CVS, doesn't it resurrect=20
> the file from the Attic and begin adding new revisions to that=20
> resurrected file?
> How might this state come to exist? Did an older version of CVS not=20
> resurrect files from the Attic?

Chances are it's a bug in CVS. It has had issues with bringing files back=
to life. I've specifically seen a different issue where the file in the=20
Attic becomes the live one, which causes no end of troubles.

> What is the correct interpretation of these two files? Is there one=20
> revisions 1.1 which is "authoritative" in this case, and one which=20
> "should" be ignored, or does the data in the Attic/Makefile,v have any=20
> further meaning?
> These kinds of inconsistencies are all over the place. Do they matter?
> Also, there are a pile of branch tags scattered around in the CVS tree=20
> that don't have any associated RCS revisions. How should that=20
> information be interpreted? For example:

Those revisions were probably zapped as part of cleaning the tree in=20
response to the AT&T/UCB Lawsuit. We retained commit comments but deleted=
diffs for "tainted" code.

That's why we are now able to have CVSweb.

Take care,


Content-Type: application/pgp-signature
Content-Disposition: inline

Version: GnuPG v1.2.3 (NetBSD)