On 11 Nov 2005, at 2:32 pm, Junichi Uekawa wrote:
Hi,CVSsuck is a mirroring tool for CVS repositories. Unlike other tools such as CVSup or rsync, it uses cvs command to access the repository. So, it works well with remote repositories without a special server or shell account. However it is inefficient and not perfect because CVS client/server protocol is not designed for mirroring. If a server provides special way to grab a repository, you shouldn't use CVSsuck.Sounds like this is best kept out of the archive then...To me it sounded like a pretty good idea, if the server only provides anonymous pserver CVS access.
I wouldn't call it a "mirror" though; how does it manage to fetch the complete repository including history? It doesn't do something evil like fetch the cvs log, and then fetch every single revision for every file, does it? I shudder to think what would happen if it does that and someone pointed it at the largest public CVS server I maintain (cvs.sanger.ac.uk) which has several GB of CVS repository. The Ensembl source code alone has had about 60,000 revisions made to it in its life (judging by the revision number that I ended up with when I converted it to a Subversion repository as an experiment).
If it's not doing anything as complex as that, what does it give you that just using cvs checkout doesn't?
Tim