Improve architecture for Model updates
Some issues make model update work intensive and should be improved:
- Vaadin and TaxEditor still have their own dataset.dtd. They replace it by dataset.dtd from cdmlib-test (maybe same for 001-cdm.h2.sql
- Tests in Vaadin and/or Editor are too big. Many redundant test data.
- TaxEdtior (also Vaadin??) has its own hibernate.cfg.xml, should be replaced by cdmlib-test or cdmlib-persistence hibernate.cfg.xml
- CdmModelCacher should be moved to cdmlib and should be started automatically with each build, stored either in model, persistence, service or io (?); maybe used by cdmlib itself; also it needs some dependency to lucene etc which doesn't exist in TaxEditor. We could also use a simplified configuration without lucene(?).
- 001-cdm.h2.sql should be created automatically (there is still an issue with the envers part) => #6714
- ClearDB_with_Terms_DataSet.xml, ClearDBDataSet.xml and TermsDataSet.xml is still required in each project. Should only be in cdmlib-test.