Today I ran across my first majorly annoying issue with using Subversion for source code control.  It appears to automatically add hidden .svn directories all up and down the checked out source code tree.  Not sure exactly why that approach was chosen, my guess is that it makes it simpler to handle updates/commits.  Unfortunately, it becomes a huge hassle when trying to slide a repository from one machine to another.  It appears that you've either got to manually edit a file in each .svn directory to point to the new server, on each user's machine, or you have to get ALL users to stabilize and commit all their code, so that they can delete local copies and then re-checkout from the new server.  There's import/export support in subversion, and we used that originally to take a single user source and push it to our first server.  Unfortunately, it doesn't seem to help at all in this situation.  Very annoying, especially since the repository itself can very easily be ported with a simple tar/sftp/untar operation.  It's all those end-user copies with the hidden directories (and hard-coded, upon checkout, info) that makes it painful.