StatCVS is a great program - I used it for many years as a USAF contractor. However it has a couple of limitations:

• it doesn't compile stats from CVS branches
• it takes a REALLY long time to run (mostly because it requests each version individually from the server)

This implementation fixes both of these issues (the latter by reading the comma-v files directly). It's probably pretty buggy, and could use some definite refactoring, but it works well for me and my current employer.

Additionally, it links to Bugzilla bugs specified in CVS comments. And links to a ViewVC repo for diffs (if you have one, which I would recommend).