Niclas Hedhman wrote: > Ok, we are not there yet. But basically the runtime will detect schema > changes "per-object' (in reality per state of object) and when a > mismatch between "running" and "stored" schema is detected, the > UnitOfWork will look up a 'migration-path'-hook which you install. > Exactly how this is going to look is still debatable, and we welcome > any insights. > > The next step would be to enable the 'hooks' to be automatically > created for common refactorings and changes to the code, such as > simple renames, moves, deletes and additions, from within the IDEs > refactoring system. Sounds too 'automagic'?? Yeah, maybe, it remains > to be seen.
This needs to be done either on per-object, or more preferably, per-aggregate, level. Some of the main hooks are already there, such as schema version calculation. Now we need some ideas on how to actually do it. In the meantime i think you can basically use the REST interface to get the raw data, transform it, and put it back in with the new format. Not tested though. I need this myself for StreamFlow really soon, so it's going to need some attention. /Rickard _______________________________________________ qi4j-dev mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/qi4j-dev
