On Mon, Sep 1, 2008 at 12:16 PM, Rickard Öberg <[EMAIL PROTECTED]> wrote:
> Right, gotcha. There's nothing that says it can't work, but there's also > nothing that explicitly says that it must work. I'm not really sure what > the best way to do it is. As Niclas pointed out in the issue text, it > may open a whole can of worms... > > Would it be possible to do a test and see what happens? The main problem I see is that Property works very well for 'native types', whereas it may seem not suited for complex values. For instance, if a Property can have a complex type, then it may consist references to Entity, Value and Transient composites. How do we deal with each of those cases? How about Property of a complex type with Associations? Somehow, I feel that either Property should only allow 'native types', or that it additionally should allow ValueComposites since we probably have to deal with similar issues in the ValueComposite semantics definition, and all other types are not allowed. Exactly what constitutes a 'native type' is of course up for debate, but perhaps a Serializable (without special handling) would do just fine. Cheers Niclas _______________________________________________ qi4j-dev mailing list [email protected] http://lists.ops4j.org/mailman/listinfo/qi4j-dev
