sql.py: put_table(): Pass pkeys to run_query_into() by ref so it will be updated if the table name is changed

sql.py: run_query_into(): If CREATE TABLE AS generates a DuplicateTableException, rename the table with a version # prepended

sql.py: run_query_into(): Made into param a reference so that the function can change it, and renamed it to into_ref

sql.py: run_query_into(): Made into param a reference so that the function can change it, and renamed it to into_ref

sql.py: put_table(): If DuplicateKeyException: run_query_into() recoverably, so that DB errors such as DuplicateTableException will be parsed

sql.py: Removed no-longer-needed try_insert()

sql.py: Merged with_parsed_errors() into run_query() so all recoverable queries would automatically benefit from DB error message parsing. DbConn: Moved _add_cursor_info() to DbCursor.execute().

sql.py: with_parsed_errors(): Raise DuplicateTableException for "relation already exists" errors instead of "table name specified more than once" errors

sql.py: run_query_into(): Removed "DROP TABLE IF EXISTS" because sometimes when there are collisions in the temp table names, the code actually uses both "copies" of the temp table. Eventually, this situation will be resolved by adding a counter to the temp table name.

sql.py: Cleaned up DbException's and subclasses' messages