Hi Jannis.

About a year ago, I too became frustrated by the lack of a good statistics tool for Subversion, as I had used StatCVS-XML for all my CVS projects with other people. And like all programmers with too much time on their hands, I created my own tool which I called SubStats - notice the missing "v", Cyrus! :)

It has some funky features like being completely character-based (ignoring white-spaces) such that an author is not praised by how many *lines* he or she writes, but rather how many actual characters are written. It is also able to check for moved content within the same file. Even more funky, you can specify that certain files should not be included in the statistics initially, such that an author can commit a file he or she has not written, and it will not influence the statistics, but any subsequent changes to that file will. The code is one big pile of spaghetti though.

I have only recently (a few days ago) released it on the web. It is still a bit wet behind the ears, so I don't expect it to work perfectly yet. Also, it does not yet have all the functionality from StatCVS, and it is only tested on Linux. However, if you can get it to run, I am sure it can be very useful.

It is written in Ruby, but I have packaged it as a Linux binary so you don't actually need a Ruby interpretor.

You can find it here, along with a report sample: http://www.molgard.eu/substats/.