Well, it does define ResponseBean that how it would look like and that
the additional parameter would be AssynHandler<ResponseBean>.
The only case when the ResponseBean doesnt appear in the java signature
is when the output message has only 1 part. In this case the
corresponding boxed java mapping of the output part would appear in as T
in AsyncHandler<T>.
See the last para of 2.3.4.4:
"If the resulting response bean has only a single property then the bean
wrapper should be discarded in method 28
signatures. In this case, if the property is a Java primitive type then
it is boxed using the Java wrapper type 29
(e.g. int to Integer) to enable its use with Response."
-vivek.
Kohsuke Kawaguchi wrote:
>
> Thanks. But my understanding is that that part explains how does a
> response bean look like, not about how the async method signature
> looks like. Is that correct?
>
> Vivek Pandey wrote:
>> I am saying based on my understanding of sec 2.3.4.4, where it
>> defines the rules as to what would be the resposne bean:
>>
>> ? If the operation’s output message contains a single part and that
>> part refers to a global element decla- 20
>> ration then use the referenced global element. 21
>> ? Synthesize a global element declaration of a complex type defined
>> using the xsd:sequence com- 22
>> positor. Each output message part is mapped to a child of the
>> synthesized element as follows: 23
>> ? Each global element referred to by an output part is added as a
>> child of the sequence. 24
>> ? Each part that refers to a type is added as a child of the sequence
>> by creating an element in no 25
>> namespace whose localname is the value of the name attribute of the
>> wsdl:part element and 26
>> whose type is the value of the type attribute of the wsdl:part element
>>
>> the specific example you gave as 2 output parts so both of them
>> follow the second bullet where the output parts appear inside the
>> response bean. The response bean is named after wsdl:operation_at_name +
>> "Response".
>>
>> -vivek.
>>
>> Kohsuke Kawaguchi wrote:
>>> Vivek Pandey wrote:
>>>>> Is this the correct signature to generate? If so, I need to do
>>>>> some surgery on client SEIStub.
>>>>>
>>>> Yes, the generated signature looks correct. The response parameters
>>>> (headers and body) appear as property in the response bean.
>>>
>>> But on what basis do you say that the generated signature is correct?
>>> I'm not seeing a relevant description in the spec.
>>>
>>
>
>
--
Vivek Pandey
Web Services Standards and Technologies
Sun Microsystems Inc.