I'm sorry, forget this task. I checked out the JDBC spec and it seems that it's quite
buggy from my point of view! The spec
tell that if you do a SQL-to-TYPE mapping, SMALLINT will map to type short, but if you
do a SQL-to-CLASS mapping (part 8.14
of JDBCTM Guide: Getting Started) a driver has to map SMALLINT to java.lang.Integer.
My god, why that?
This means that there have to be NO changes to GenIC or JOnAS, but that everyone that
has a bean with short fields, and now
wants to hold the NULL information (and therefor changes from TYPE to CLASS mapping),
short becomes not Short but Integer.
Strange, but true!
