Apache Derby support in MCF has two major problems. The first problem is that is not upgradeable, which means that people who inadvertantly "go live" using Derby as MCF's underlying database have no place to go. The second problem is that it is unstable (see CONNECTORS-998). The second problem would not be so bad if it wasn't for the fact that Derby bug fixes are few and far between these days. The maintainers of that project don't seem to have a lot of internal knowledge of the offering anymore.
Since we have a much better alternative available and fully supported, namely HSQLDB, I am proposing that we remove Derby support entirely in MCF 2.0. This will not only reduce our bin and lib distribution size significantly, but our tests etc. will run faster and be more reliable. The work involved is mainly around transitioning existing tests that run only on Derby to run on HSQLDB instead, as well as the single-process example. It's a major amount of work, but I would expect it to be straightforward. Thoughts? Karl
