I got a new version from CVS yesterday, and I saw two new changes breaking 
things for us that worked in the last version I checked out (Feb 6). I have 
tried to track this to changes made to the code announced to this mailing 
list without any success, so I'd appreciate if someone code shed some light 
on why these changes were made, and possible workarounds.

Issue 1: shared TypeMappings
---
We have a lot of BeanMappings that must be shared among multiple services, 
and they are hence put in the deployment scope of the server configuration, 
and not within the local service element. In the new version the shared 
TypeMappings are not picked up by the engine. The only mappings to be 
picked up are the ones specified in the local service tag in the 
deployment. This can be easily reproduced by modifying the deployment 
descriptor of samples.userguide.example5 to move the BeanMapping out to the 
global scope.

Issue 2: BeanSerialization of java.lang.Object (xsd:anyType)
---
We have a lot of extensibility elements in our WSDL that are of type 
xsd:anyType. When we send these across the wire on their own we see the 
expected behavior, i.e. the client sends off a request with xsi:type set to 
the actual type, and the server maps the type to the correct deserializer.
However, when we  send a complexType containing an xsd:anyType element 
using the BeanSerializer, the client sets the type to xsd:anyType, and the 
server is unable to deserialize the request (this happens even if a xsi:nil 
is set) with the error: org.xml.sax.SAXException: !! No Deserializer for 
class java.lang.Object.

Both of these issues are showstoppers at the moment, so I'd appreciate any 
pointers, e.g. on how to modify/customize the BeanSerializer, and 
TypeMapping registries to get it to behave like the old version (checked 
out on Feb 6) when we didn't have any of these problems.

Thanks,
     Thomas


Thomas Sandholm <[EMAIL PROTECTED]>
The Globus Project(tm) <http://www.globus.org>
Distributed Systems Laboratory
Mathematics and Computer Science Division
Argonne National Laboratory

Reply via email to