(no subject)

there are automated cvs->subversion migration tools, called cvs2svn (at least the debian package).

if you like to continue to use s.th. like svn+ssh://repopath/revision you should add wrapper scripts arround the svn server scripts, which prevent them to change the owners on the database, which leads to the fact only one person can work with the repository.

If you use a verry recent version of svn, i would suggest you to use file based storage instead of the bdb one, because of the bdb storage tends to mess up on a frequent base.

I personaly prefer to have a _seperate_ tree for every svn, to not make them cross interfere each other with db corruption etc. (which is, again, a serious issue with the bdb backend)

This has the drawback, that you need to configure every repository in apache seperately with a <location /svn/project> and have to restart apache.

 There are allso some neat ways to screw your checked out repo...

Prepare to have more problems than you had while getting into CVS. (the bsd'ish  "no clue" issue ;-)

But once it works, it's fine.

I've found a neat frontend to it over at opensync.org called 'TRAC', which offers wiki and svn and time management.

seems to be written in python (horay!;-)