On Wed, Mar 25, 2009 at 11:59 AM, Rickard Öberg <[email protected]> wrote: > Niclas Hedhman wrote: >> >> Yes, this requires some thinking. >> >> I think you also realize that this is not a problem for REST URLs >> only. As you pointed out, EntityStore.getEntityState() takes the type >> as input and if it has been refactored, it is 'gone'... > > Something like that, yeah. It might be that getEntityState() should simply > take id as input, and whatever it was stored with should be the type, and > not what the client suggests it is.
The ID passed at that level is an opaque String? Would that create problem for 'mapping'-like stores? >> It also kind of feels odd that we need to involve the EntityStore in >> the resolution of schema version. What happens if "type" is not known >> to the EntityStore at all? What is affected and currently needs the >> type information? > > Exactly. Could we put EntityTypes and schema handling in a separate service > perhaps? Yes, I think that would make sense. In fact, UoW is fairly overloaded and perhaps there are other things it does that perhaps should be broken out as well. Cheers Niclas -- http://www.qi4j.org - New Energy for Java _______________________________________________ qi4j-dev mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/qi4j-dev
