+1, I think that makes sense. Nevertheless, if I recall correctly, both layers make use of non-standard JPA features, that might raise some issues.
Thanks, Tammo Milinda Pathirage wrote: > +1 for the migration and separate out JPA implementation specific stuff in > trunk. Also we should consider about improvements to JPA based > implementation. > > Thanks > Milinda > > On Wed, Nov 18, 2009 at 2:25 AM, Greg Lucas <[email protected]> wrote: > >> +1, that sounds cleaner. >> >> >> >> On Mon, 16 Nov 2009 03:12:31 -0500, Rafal Rusin <[email protected]> >> wrote: >> >> +1 from me for migrating OpenJPA to JPA (supporting OpenJPA + >>> Hibernate JPA) + dropping dao-hib in ODE trunk. >>> >>> 2009/11/16 Jeff Yu <[email protected]>: >>> >>>> Hi all, >>>> >>>> I looked into our Ode DAO impl , it has three implementations at the >>>> moment, >>>> one is Hibernate, one is Memdao, which I thought it is meant to be used >>>> in >>>> test environment, the other one is JPA, but it ties to openJPA somehow, >>>> as >>>> it uses the some OpenJPA API directly in the implementation. >>>> >>>> Searched the 'hibernate openjpa' keyword in our maillist, only found this >>>> thread(http://ode.markmail.org/message/mjqsz6ehykpzbons), which is back >>>> in >>>> Dec, 2008. Little info about these info. Well, I am trying to deploy Ode >>>> with hibernate as JPA provider. But I learnt from an email that we may >>>> drop >>>> hibernate implementation in the future, is this still the case? If so, I >>>> guess we won't provide the migration tool for users from 1.3.x to 2.x >>>> version. Will we provide a migration tool from the hibernate -> openjpa? >>>> >>>> One thing that I didn't fully understand is that we've already tried to >>>> use >>>> the JPA as our second level of abstraction for the DAO, (the first level >>>> abstraction I see is our DAO interface), but I see we tied this impl into >>>> OpenJPA by using following classes. >>>> >>>> import org.apache.openjpa.ee.ManagedRuntime; >>>> import org.apache.openjpa.util.GeneralException; >>>> >>>> import org.apache.openjpa.persistence.OpenJPAPersistence; >>>> import org.apache.openjpa.persistence.OpenJPAQuery; >>>> >>>> It would be great that we can abstract these classes out from our jpa >>>> module, so that openjpa and hibernate can have its own impl. I understand >>>> the generated schema is pretty different in the openjpa and hibernate, >>>> but >>>> they are transparent to users/developers. I meant, as long as they are >>>> using >>>> the right combination (Hibernate + hibernate generated schema, or OpenJPA >>>> + >>>> OpenJPA generated schema), they should be fine. >>>> >>>> I've noticed that in the hibernate implementation, it is using the >>>> hibernate >>>> specifc annotation, which seems to me that it hasn't been actively >>>> developed >>>> for a long time. >>>> >>>> I'd like to see us using the JPA in its impl, and can be deployed in any >>>> JPA >>>> provider, in this case, it would be easy for users to deploy ode into any >>>> JEE application server. I didn't know whats our plan for our DAO impl, >>>> both >>>> for the 1.3.x and 2.0 version. Also, some background on this and some >>>> issues that we've had in the implementation would be of great helpful to >>>> me. >>>> > >
