develooper Front page | perl.perl6.source.control | Postings from September 2000

Re: Can perforce gateway to CVS without loss of metadata? (was Re: code repository)

Thread Previous | Thread Next
Adam Turoff
September 8, 2000 10:35
Re: Can perforce gateway to CVS without loss of metadata? (was Re: code repository)
Message ID:
On Fri, Sep 08, 2000 at 11:30:38AM -0400, Bradley M. Kuhn wrote:
> Adam Turoff wrote:
> > *: Sarathy tells me that Perforce sucks at maintaining thousands of
> > anonymous checkouts, while CVS doesn't mind at all.  This is a perfect
> > reason to use anon CVS vs. Perforce, but does not require that Perforce be
> > ditched in favor of CVS, only that an anon CVS gateway be maintained.
> Will this anonymous gateway have all the relevant meta-data?  Will it look,
> to a user of the gateway, like the live repository is CVS when they go
> hunting through metadata?

Since there is no anonymous gateway, the answer is a superposition of:

	1) Yes, that's the whole point
	2) No, it doesn't exist yet
	3) Maybe, if it's implemented properly
	4) Maybe not, since Perforce is differently featureful


Sarathy and I talked about the possibility of setting something up 
that was automated.  If there's a frequent rsync with commits into
a parallel CVS repository, then the first (and most important thing)
that is lost is who actually made any specific update; all changes
are made by 'cvs-bot' or somesuch.

Other approaches have been talked about, like Simon's Perforce-like
front end to CVS, and Barrie's generalized interface to 
Perforce/CVS/BitKeeper that should be able to pretend a Perforce
repository looks like CVS to a CVS client.

The first approach is easiest to implement and only gets at the
spirit of the problem - anonymous CVS access to the tree.  Tracking
the Perforce change numbers shouldn't be a huge issue, but losing
ownership of who's changing the source (via 'cvs annotate') is a
pretty big loss.  In all fairness, I haven't spent many cycles trying
to figure out how to fix that.  If we're talking about ~30 committers
or less, it shouldn't be too horrific.


Thread Previous | Thread Next Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at | Group listing | About