Now, with the binding, I have already try a lot of combinations using
extends, abstract, and others (even the create-type for the
Class A binding to create Class B instead)... but, when unmarshalling,
only one class is created with only and ONLY his particular atributes.
Here is the basic
Hello,
The type of field person in class Customer is of type Person. Therefore the
mapping compiler is able to map to the right class. Even so I use the map-as
attribute to document the assoziation within the mapping file.
Mathias
Original-Nachricht
Datum: Thu, 7 Sep 2006
Hello,
if you want to use A as an instance as your mapping suggests you cannot make A
abstract. The extends relation is not needed in your context.
Here the snipped that should work:
mapping name=classA class=mypackage.A flexible=true odered=false
value name=go field=pepe/
value name=two
Hmmm, you too? :-) I've been struggling with this for the last few days. There seems to be no consistency. If you use 1.1 for runtime and 1.0 for compile, it will complain, etc...To resolve, go to your local repo and change the pom file for maven-plugin to require
1.1 for both runtime and binding
Hi again, I resolved (in some way) my problem of extending classes
using a combination of abstract an extends; but now I can not make the
abstract mapping being ordered=false and flexible=true (and, trust
me, I need it).
When doing that only the fields of the base mapping are being
unmarshalled,
I am currently evaluating JiBX for the use of replacing our current
serialization to a blob. We have about 75 classes that form a
complicated graph (with numerous circular references, and classes
inheriting from other classes, some of which are abstract) and are
currently using Java's default