[386] in NetBSD-Development
Re: Problems with /afs/sipb/project/sipb-athena
daemon@ATHENA.MIT.EDU (Ken Raeburn)
Fri Jan 13 14:09:39 1995
From: Ken Raeburn <raeburn@cygnus.com>
Date: Fri, 13 Jan 1995 14:09:09 -0500
To: jhawk@MIT.EDU
Cc: netbsd-dev@MIT.EDU
In-Reply-To: jhawk@MIT.EDU's message of Fri, 13 Jan 95 13:29:11 -0500 <9501131829.AA01808@hodge-podge.MIT.EDU>
From: jhawk@MIT.EDU
Date: Fri, 13 Jan 95 13:29:11 -0500
> What we really need is something like CVS or ClearCase so we can share
> the Athena source tree and do our local hacking and not lose track of
> each others' stuff.
It's not clear to me that CVS will help us with the Athena source
tree problem. There are multiple Athena source trees, and it's not
clear we want to keep track of development in all of them.
If you set up your source repository, note in the log where a package
comes from when you import it. Or, if you get Athena to start using
CVS for everything, or for individual projects (e.g., Zephyr?), CVS
will automatically keep track of where a directory was checked out
from.
I expect ClearCase would let you do the same... Naturally, I'd vote
for cvs because it's a free solution, but I'm not actually working
with your sources...
It's also not clear to me that we have the disk space to deal with
CVS in a convenient fashion (I've said this before. I'd be happy
to be wrong).
Modify CVS to use gzip on the repository.
Use symlinks to Athena sources when you can.
The main problem I see with using CVS is that there are different
versions floating around. There're the regular net versions, which
are rather broken in how they handle file deletions, and the Cygnus
version, which does it coherently but requires a special version of
RCS. On the other hand, I'd always recommend picking up and building
a recent version of RCS and forcing CVS to use it, just because so
many vendors supply ancient versions.
So maybe the Cygnus version would be the way to go... I think that's
what Kerberos 5 is using.
