Date: Thu, 03 Jul 1997 17:59:38 -0700 From: John Polstra <jdp@polstra.com> To: Bruce Evans <bde@zeta.org.au> Cc: ache@nagual.pp.ru, current@FreeBSD.ORG, xaa@stack.nl, peter@spinner.dialix.com.au Subject: Re: CVSUP and CVS branches: still not synchronized, please fix! Message-ID: <199707040059.RAA20399@austin.polstra.com> In-Reply-To: Your message of "Fri, 04 Jul 1997 04:39:22 %2B1000." <199707031839.EAA26211@godzilla.zeta.org.au> References: <199707031839.EAA26211@godzilla.zeta.org.au>
next in thread | previous in thread | raw e-mail | index | archive | help
> Perhaps the cvs bug that caused histedit.h to reappear in all
> checked out trees only affects cvsup for disklabel.5.5.
I think the problem is a botch in the CVS repository. (Hi, Peter!)
RCS file: disklabel.5.5,v
Working file: disklabel.5.5
head: 1.1
branch: 1.1.1
locks: strict
access list:
symbolic names:
bsd_44_lite_2: 1.1.1.3
bsd_44_lite: 1.1.1.1
CSRG: 1.1.1
keyword substitution: kv
total revisions: 4; selected revisions: 4
description:
----------------------------
revision 1.1
date: 1994/05/26 06:34:00; author: rgrimes; state: Exp;
branches: 1.1.1;
Initial revision
----------------------------
revision 1.1.1.3
date: 1997/07/02 00:24:14; author: bde; state: Exp; lines: +6 -2
Import Lite2's src/sbin, except for XNSrouted and routed. All relevant
files in src/sbin are off the vendor branch, so this doesn't change the
active versions.
----------------------------
revision 1.1.1.2
date: 1995/12/30 18:51:57; author: peter; state: dead; lines: +0 -0
recording cvs-1.6 file death
----------------------------
revision 1.1.1.1
date: 1994/05/26 06:34:01; author: rgrimes; state: Exp; lines: +0 -0
BSD 4.4 Lite sbin Sources
Note: XNSrouted and routed NOT imported here, they shall be imported with
usr.sbin.
----------------------------
Notice that the death of the file was recorded in 1.1.1.2, which
is on the vendor branch. This should never happen. CVS will never
do this. It must have been done by hand with raw RCS commands.
CVS would have recorded the death on the main branch, in revision
1.2. If that had been the case, then Bruce's subsequent vendor
branch import would have come up as a conflict, and the merge would
have again left the file dead.
Since the death was incorrectly recorded on the vendor branch, the
new import of 1.1.1.3 brought the file back to life again as far as
CVSup is concerned. (It has "state: Exp" instead of "state: dead".)
If CVS still treats it as dead, that can only be because of some
sort of provision for compatibility with very old repositories
(pre-1.6 of CVS). I am not really interested in supporting
compatibility that far back. Anybody who objects is respectfully
invited to spend about 8 hours browsing the CVS sources, trying to
figure out how it ought to work for all cases. See how you feel
about it after that. :-)
I think the right thing to do at this point is to simply "cvs rm"
the offending file again.
John
--
John Polstra jdp@polstra.com
John D. Polstra & Co., Inc. Seattle, Washington USA
"Self-knowledge is always bad news." -- John Barth
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199707040059.RAA20399>