I don't see any problem with the spec currently wrt this. Of course I may find something the deeper I get but for now my problem is that the bval implementation uses the the class and the POJO methods directly rather than some plugable way to determine these things.
On Thu, Apr 16, 2015 at 3:58 PM, Mark Struberg <[email protected]> wrote: > I think we probably could experiment with such things at the bval impl > level. But I doubt that this is possible by utilizing the spec API. > > LieGrue, > strub > > > > Am 08.04.2015 um 17:31 schrieb Keith Garry Boyce <[email protected]>: > > > > I was looking into using bval to do RDF (triple store) validation but > the API currently does not allow for modifications at that level.. > > What I want is to have a validation xml like follows: > > > > <bean class=" > http://cambridgesemantics.com/ontologies/2009/08/Predicate#Person"; > ignore-annotations="false"> > > <field name=" > http://cambridgesemantics.com/ontologies/2009/08/Predicate#reqProp";> > > </field> > > </bean> > > > > Where instead of the class being a class it's at URI defining an RDF > type and the fieled is a URI defining a predicate. > > > > When I follow the code base in ApacheValidatorFactory it instantiates > new ValidationMappingParser > > > > > > which appears to be very class centric hence I can't really implement a > variation that doesn't just override the accessors. > > > > I would like to be able to do validate(aPerson) and in configuration > define what property off the object defines the type rather than the Class > itself. > > Then I would also like to overide the field accessors so I can make my > own implementation of how to extract the value of each field from the > object. > > > > So simply I'm asking for the standard java validation spec (or at least > this implementation) to allow for more flexible validation options. > > > > Any thoughts? > >
