I agree that java.lang.Object is what we want, and yes, a generated SDO loader will populate POJOs, whether or not they extend from DataObject.
But, if I understand Raymond's comment, I think he's trying support the case where we don't generate a loader, of maybe not even a class at all, but instead use Dynamic SDO. Those cases would require integration with a loader that works with DataObjects. But again, that should be no problem if the return type restriction is relaxed to java.lang.Object. Frank. Jeremy Boynes <[EMAIL PROTECTED]> wrote on 06/12/2006 04:10:09 PM: > Raymond Feng wrote: > > > > Please note the return type is required to extend from ModelObject. > > Can it be relaxed to java.lang.Object so that loader implementations > > can return their own types directly? > > > > Yes - this fits with the idea that people can use any POJOs as model > objects and are not tied to a specific framework. > > > I'm trying to add some generic loaders backed by XML databinding > > frameworks such as SDO. These frameworks usually have their own model > > objects, for example, SDO uses commonj.sdo.DataObject. It's not very > > nice that we have to wrap a native model object into our ModelObject. > > > > > > I think you have it backwards here. It is the model objects that are the > native form, not the artifacts generated by some XML binding framework. > Don't forget that Frank is in the process of enhancing SDO to add the > ability to load POJOs directly. > > -- > Jeremy > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
