>>>> I have a simple test method:
>>>>
>>>> Long foo(Long bar)
>>>>
>>>> Java2WSDL maps the Long to xsd:long and the client stub contains the
>>>> method as:
>>>>
>>>> long foo(long bar)
>>>>
>>>>
>>>
>>> How does the WSDL look like ? The parameter and return value should
>>> have a "nillable='true'" attribute, otherwise I guess it's a bug in
>>> the WSDL-generation part (Java2WSDL).
>>> If it has this attribute, then the bug is in the client-side code
>>> (WSDL2Java).
>>>
>>
>> For a method:
>>
>> Long b(Long id) throws CustomException { ... }
>>
>> Here's the WSDL snippet generated by Java2WSDL:
>>
>> <wsdl:message name="bRequest">
>> <wsdl:part name="in0" type="xsd:long"/>
>> </wsdl:message>
>>
>> <wsdl:message name="bResponse">
>> <wsdl:part name="bReturn" type="xsd:long"/>
>> </wsdl:message>
>>
>> So Long is mapped to xsd:long but it's NOT declaraed nillable. Is this
>> the bug?
>
>
> I think so. Which version of Axis do you use ?
Axis 1.2RC1
>
>>
>> Or should the Long have been mapped to soapenc:long instead which is
>> not compatible with .NET?
>
>
> xsd:long and soapenc:long should be the same thing (IIRC the soap
> encoding spec defines soapenc:long as xsd:long).
>
>>
>> Is xsd:long nillable=true compatible with .NET?
>
>
> I don't know :) It should but...
I'll just use the primitive "long", at least I know that works ...
If anyone else has an example on how to get the Long stuff to work
end-to-end and to interoperate with .NET I would love to see it (it has
to be RPC not doc/lit) as it would be nice to make use of the nillable
property ...
Thanks.
Tim
