I dont understand why scm should exclusively choose either svn or cvs.
Why cant it get from both.

I use svn but I need a module that another department manages with cvs.

So, for any simple parameterized build (I choose dev/qa/production) I need:

• job1 that collects the parameters and serializes them (with code I keep in svn) for job3 to use.
• job2 to get from cvs.
• job3 to get from svn, deserialize the parameters from job1 and do the builds.

This is counterintuitive and the whole serialization/deserialization is useless (it has nothing to do with the build itself)