At Fri, 6 Nov 2009 13:36:35 -0500, Thor Lancelot Simon
<tls%panix.com@localhost> wrote:
Subject: Re: git copies of cvs modules available
>
> I'm not sure how a ,v file which works properly with CVS can be called
> "corrupted".
If I understand correctly CVS is _not_ "properly" interpreting the
corrupted ,v file either because it is in fact corrupted, perhaps by
being structured in such a way that no possible sense can be made of the
revisions and branches given the "proper" interpretation of such things
in the context of CVS (as opposed to RCS -- i.e. the file might be a
legal and proper RCS file, but not a legal and proper CVS file).
For example with src/games/hunt/Makefile there's some really bizarre
confusion w.r.t. the vendor branch -- it is in an impossible state.
(revision 1.1 and revision 1.1.1.1 MUST always be identical!)
I suspect there are several other cases where vendor branch abuse in the
NetBSD CVS repository will cause major problems for _any_ conversion tool.
There may also be major problems in the NetBSD CVS repository caused by
the CVS-incompatible way NetBSD developers have been instructed to move
files.
RCS file: /cvs/master/m-NetBSD/main/src/games/hunt/Makefile,v
Working file: Makefile
head: 1.1
branch:
locks: strict
access list:
symbolic names:
[[ ...elided... ]]
keyword substitution: kv
total revisions: 2; selected revisions: 2
description:
----------------------------
revision 1.1
date: 1997/10/04 09:11:14; author: mrg; state: Exp;
branches: 1.1.1;
build hunt.
----------------------------
revision 1.1.1.1
date: 2003/06/11 11:45:14; author: wiz; state: Exp; lines: +204 -3
Import of hunt as of 2003-04-16 from ftp://ftp.cgl.ucsf.edu/pub/hunt.shar.Z,
with files moved to proper subdirs.
Changes since 19930717: BSD licensed, and minimal fixes.
=============================================================================
--
Greg A. Woods
Planix, Inc.
<woods%planix.com@localhost> +1 416 218 0099 http://www.planix.com/
Attachment:
pgpJRAIstFbZ7.pgp
Description: PGP signature