1. Added a check that PG was initdb's with the collation order for
characters that we expect.
2. Jon added a "sql/common" directory for the insert statements which
load data for acs-references, so I added a new datamodel type "common"
to the APM.
3. Created a new db API routine that returns true if a given db_type is
compatible with the current RDBMS.
4. Added a workaround for a PL/pgSQL "execute" bug in acs-workflow (other
cases had been fixed previously by others but this one was missed)
5. Diddled a couple of .info files using the APM.