Subject: Re: NOTICE: Weekly tar_files changing
To: NetBSD-current Discussion List <firstname.lastname@example.org>
From: Greg A. Woods <email@example.com>
Date: 10/28/1998 23:45:58
[ On Wed, October 28, 1998 at 17:33:22 (-0800), Jason Thorpe wrote: ]
> Subject: Re: NOTICE: Weekly tar_files changing
> On Wed, 28 Oct 1998 19:50:56 -0500 (EST)
> firstname.lastname@example.org (Greg A. Woods) wrote:
> > Until/unless there's anonymous CVS access to the NetBSD repository this
> > "feature" is useful only to a tiny fraction of NetBSD users, and a
> > hindrance to everyone else who uses a local CVS repository.
> (1) "Why do you think we're making this change now", and
Well, this fact could have mentioned in the same message. I certainly
didn't want to make such a bold assumption on my own despite recent
hints in related venues. How did that thread go about commications
to/from/with "core" go again????
> (2) "CVS ignores CVS/ directories when you import anyhow".
The import's not the issue. It's the potential to clobber existing
working directories. Such clobbering cannot ever possibly be reversed
and can totally destory the usefulness of an existing working
directory. Say for instance I wanted to extract just one or two
directories from the tar into my own local working directory (as
derrived from my own local repository) but I neglected to prevent the
contents of the CVS directories from being extracted. I would then have
to completely clear out those directories and re-created them to restore
the integrity of my working directory.
IMNSHO if somone want's a working directory they should just check it
out. If they want a source tree they should not be encumbered with CVS
I don't think the benefit of being able to create a working directory
from a tarball is greater than the danger of someone accidentally
clobbering their own local working directory.
Of course if CVSup access were provided instead of anonymous direct CVS
access this would be even less of an issue since remote users wouldn't
put inordinate demands on the CVS server when they went to check out a
Note that unless these foreign working directories are created using the
exact same command a user might use to access the annonymous CVS
repository then the content of the CVS sub-directories is somewhat
useless anyway because the recent versions of CVS are (finally) losing
their ability to edit the CVS/Root file on the fly and such edits would
have to be done manually by the user.
The real bummer about this is that this implies that the release images
aren't being created with 'cvs export' as per the recommended release
management procedures in the CVS manual (which would guarantee 100% that
only the tagged revisions of the sources could be included in the release).
Greg A. Woods
+1 416 218-0098 VE3TCP <email@example.com> <robohack!woods>
Planix, Inc. <firstname.lastname@example.org>; Secrets of the Weird <email@example.com>