Integrate repoze.evolution into Karl

[Chris Rossi]

We should go ahead and integrate repoze.evolution into Karl so we don't have to remember to manually run misc update scripts to evolve the db in parallel with code.

[Chris Rossi (chris-archimedeanco)]

Chris, do you have time for some more minor Karl tweaking?

[Chris McDonough (chrism-plope)]

I have:

- Made the "osi" package depend on repoze.evolution

- Created an "evolve" console script that is generated by the OSI package

To add an evolution step:

- create a file in the "osi" package's "evolve/zodb" directory named "evolveX.py" where X is a monotonically increasing number across time.

- change the value of "VERSION" inside "evolve/zodb/__init__.py" to an integer representing the "code version" (should match
  the "X" in "evolveX.py" as you add an evolve script in most cases).

To evolve a sandbox:

- ONCE FOR EACH SANDBOX: set the initial version number of the database: bin/evolve --set-db-version=X where X is the
  code version you believe the database is "already at".

- THEREAFTER (whenever evolution is required): run "bin/evolve" to see the version numbers of the code and the database.
  If they both look right, run "bin/evolve --latest" to update the database to the latest code version by running all necessary
  "evolve" scripts. Each script will be run after the other; a transaction will be committed between each script run. If a particular
  script fails due to an exception, evolution is aborted until the error is fixed.