Migrate all code to the newer clojure.jdbc API.
Note that the newer jdbc API handles some database operations like
transactions by storing bookkeeping information in nested database
references. For example, (with-db-transaction ...) creates a new
database connection map with an incremented level, and the library uses
that level to determine where the "top" is, i.e. to decide when to
To accommodate that approach (without threading a db argument through
all of the existing code), add a jdbc/*db* variable to refer to the
"current" database, and make sure to bind appropriately it in
puppetdb.jdbc wrappers for the relevant clojure.jdbc calls.
Use (jdbc/db) to refer to *db* when invoking the new jdbc API's
Use jdbc/ consistently to refer to puppetdb.jdbc in altered files.
While updating migrations, combine adjacent db-do-commands into one