Re: [Gnu-arch-users] Re: cscvs--experimental--1.1 nearing doneness; call for testers

[Andrea Arcangeli]

On Tue, Sep 30, 2003 at 03:26:15PM -0500, Charles Duffy wrote:
> On Tue, 2003-09-30 at 14:10, John Goerzen wrote:
> > I have not run a checkin on this tree since Tom added the inode cache to
> > tla, so I don't have complete stats for you yet.  I'll let you know once
> > cscvs is ready.
> 
> One initial issue here: cscvs doesn't have support for importing bkcvs
> repositories, and adding such support is not something I consider a
> priority.
> 
> Are you considering cscvs "ready" at such time as you're able to run in
> front-to-back even with the 2 redundant changesets (on average) for each
> correct one, or will you consider it "ready" at such time as it's had
> bkcvs-to-svn-like functionality?
> 
> 
> Incidentally: bkcvs makes some guarantees [every revision in a given
> changeset at the same timestamp, no two changesets at the same
> timestamp] that make it ideal for doing a perfectly accurate cscvs-style
> reconstruction very easily... but these guarantees are useless on other
> repositories, and would mean having multiple code paths in places where
> there's presently only one. If a "bkcvs mode" for cscvs is in high
> enough demand, it *could* happen -- but it's not something I'm
> particularly inclined to do; indeed, it might be better off maintained
> as a separate branch.

For a final "official" import having the bkcvs knowledge sounds
reasonable. Note that it's much easier to use the bkcvs changeset
numbers in the logs to deduce the patchsets for arch than to guess them
with an heuristic.

cvsps has been updated with bkcvs knowledge very quickly, so it
shouldn't be much more complicated for cscvs.

But overall it shouldn't make much difference either ways.

Andrea - If you prefer relying on open source software, check these links:
            rsync.kernel.org::pub/scm/linux/kernel/bkcvs/linux-2.[45]/
            http://www.cobite.com/cvsps/