«

Jan Thomä wrote:
> On 2008-08-28 17:08:55 +0200, "Mike Schilling"
> <> said:
>
>>> This would mean, that there must not be two operations having the
>>> same
>>> parameters because otherwise the service would be ambigous.
>
>>> getFolderByName(String)
>>> getFileByName(String)
>>
>> Not ambiguous if the schemas use different element names, say
>>
>> <folderName>
>>
>> vs.
>>
>> <fileName>
>
> I see, yet our server running axis 1.3 seems not to be able to
> provide
> this kind of magic and complains that there is no known operation
> called. So i stick to the --noWrap as this seems to work.

What "magic"? Write the schema that way, and generate the Java stub
and skeleton from it.

On 2008-08-26 21:56:21 +0200, "Mike Schilling"
<> said:

> Jan Thomä wrote:
>> On 2008-08-26 18:53:35 +0200, "Mike Schilling"
>> <> said:
>>> In document literal style, the body of the request is described
>>> precisely by its schema. The operatiion name is added and the
>>> parameters wrapped in rpc literal style.
>>
>>
>> Mike,
>>
>> I'm not completely sure on how to interpret your post. Are you saying
>> I did something wrong or are you saying that there is a bug in Axis?
>
> I'm saying that Axis is correctly interpreting the WSDL you've given it. If
> you want to see the method name appear in the message, use rpc/literal
> instead of document/literal.

Mike, again

could it be we are mixing up document/literal and document/literal -
wrapped style here? Because in the wrapped style, the method name is
supposed to appear inside the body of the message. In this case the
the only thing that could cause trouble in dispatching the message
would be an overloaded method. I found a nice article explaining the
differences very well:
http://www.ibm.com/developerworks/we.../ws-whichwsdl/

Btw. I have tried another WSDL using the --noWrap switch, where it
seemed not to work . I guess there are some idiosyncracies concerning
the wrapped style with axis.

Best regards,
Jan

Jan Thomä wrote:
> On 2008-08-26 21:56:21 +0200, "Mike Schilling"
> <> said:
>
>> Jan Thomä wrote:
>>> On 2008-08-26 18:53:35 +0200, "Mike Schilling"
>>> <> said:
>>>> In document literal style, the body of the request is described
>>>> precisely by its schema. The operatiion name is added and the
>>>> parameters wrapped in rpc literal style.
>>>
>>>
>>> Mike,
>>>
>>> I'm not completely sure on how to interpret your post. Are you
>>> saying I did something wrong or are you saying that there is a bug
>>> in Axis?
>>
>> I'm saying that Axis is correctly interpreting the WSDL you've
>> given
>> it. If you want to see the method name appear in the message, use
>> rpc/literal instead of document/literal.
>
> Mike, again
>
> could it be we are mixing up document/literal and document/literal -
> wrapped style here? Because in the wrapped style, the method name is
> supposed to appear inside the body of the message. In this case the
> the only thing that could cause trouble in dispatching the message
> would be an overloaded method. I found a nice article explaining the
> differences very well:
> http://www.ibm.com/developerworks/we.../ws-whichwsdl/
>
> Btw. I have tried another WSDL using the --noWrap switch, where it
> seemed not to work . I guess there are some idiosyncracies
> concerning the wrapped style with axis.

OK, I think I'd figured out where the disconnect is. Are you writing
the schemas yourself, of letting Axis generate them from Java classes?

I've been assuming the first of these. Now it looks more like it's
the second, and what you've been asking about is how to get Axis to
generate schemas that are unique for each operation in your service.
"wrapped" is neither a SOAP nor a WSDL concept, so I'm assuming it's
an Axis concept for schema generation. Yes, if you're letting Axis
generate the schemas, it looks like "wrapped" is a better choice
(though RPC/literal is probably better still, as it handles overloaded
operations more naturally.)