Implement simple JDBC implementation of `ClusteredJobDao`

[Chris Fuller]

An out-of-the-box JDBC implementation of the DAO would make Caesium more consumable as a general-purpose scheduling library.

[Chris Fuller]

I've played around with this a bit, but dealing with the various DB syntaxes, particularly for BLOB, is extremely tedious (Postgres doesn't have BLOB; some DBs prefer that you work with createBlob/setBlob/getBlob while others prefer getBinaryStream/setBinaryStream). JDBC is itself awkward.

When you consider that currently we have no products that would directly consume the JDBC interface instead of going through an abstraction layer (entity-engine for JIRA; Hibernate for everyone else) and the complexity involved in doing this correctly (meaning testing it against all the DBs as well), I just don't feel motivated to take this one on.

I've pushed up the branch abandoned/3-jdbc-dao-spike with where things were when I abandoned them due to lack of personal enthusiasm. We can revisit this later if there is public demand or if somebody wants to move my spike forward.

cc: @jhinch -- since you were the one that originally suggested doing this. ;)