> implementation of com.sun.rowset.CachedRowSetImpl is itself broken as it
> doesn't follow the JDBC specification (see
> http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7046875 ).

Wow! change with a trivial and provided patch remains posponed for year+...
Not that awful as with JAI yet still.

Maybe Java 7 team just does not know about it ? did someone tried to open that bug against JDK 7 ?
And OpenJDK teams did not fixed it, they probably are move agile than Oracle ?

Is that class covered with TCK i wonder.