Subject: Re: cvs 1.11.10 will be imported
To: NetBSD Userlevel Technical Discussion List <tech-userlevel@NetBSD.org>
From: Frederick Bruckman <fredb@immanent.net>
List: tech-userlevel
Date: 12/13/2003 14:39:10
On Sat, 13 Dec 2003, Greg A. Woods wrote:

> [ On Thursday, December 11, 2003 at 15:50:00 (-0600), Frederick Bruckman wrote: ]
> > Subject: Re: cvs 1.11.10 will be imported
> >
> > Not easy. "cvs add" balks. I had to move it out of the way first:
>
> If I'm not mistaken about what you're trying to do here, this is an
> example of one of the problems that will be encountered when trying to
> use normal branches in modules which were created with a vendor branch
> by "cvs import", especially if "cvs import" is ever used again.
>
> The whole concept of CVS vendor branches is fundamentally incompatible
> with "normal" branches.

The tests I performed had nothing to do with normal branches; there
were no normal branches. I created a top-level module containing a
single file, imported (with a vendor tag), checked out the module
(with which to observe the reported problem), moved that out of the
way, and checked out another to do the deletion of the one file. For
subsequent tests, the module was removed directly from the repository
and the process started anew.

The discussion, up to know, has been about a reproducible bug in cvs
1.11.0, that was evidently fixed (without comment!) in 1.11.5, the
relevance to the main topic being that we (NetBSD), require a high
level of confidence in another project's release process before we'll
readily import their latest release without scrutiny, if then. Now,
Greg, if you intend to carry on at length about how you feel NetBSD
misuses cvs, I suggest you start a new topic.

Frederick