First of all, thanks for your answer.
I was testing a bit and I finally managed to get Option #1 working. But I have to say that I am a bit confused and I have no idea why this is working. So I decided to take a look at MsgProvider source code. Its 'invoke' method gets a OperationDesc object from the msgCtx. And from that OperationDesc it gets a Method object ('reflection' is used to actually call the method of the service class).
But I got lost trying to find where those objects ( OperationDesc and Method ) are created and set to the msgCtx.
So it seems like several methods can be associated with a service URL like in Axis rpc/document/wrapped cases.
I will continue looking at the source code but I would appreciate if anyone has a clue about how this is working. If the Option #1 is a correct option then you don't need to extract the operation from the message. You get it for free using a method per possible operation. Then it might be a good idea to use it. But if this is unexpectably working, i.e, it is not meant to work like that, then let's forget Option #1.
The reason why I want to use Axis message style is mainly because I am using soap messages like this:
(always one only child in soap env. body)
<soap-env:Envelope xmlns:soap-env="http://schemas.xmlsoap.org/soap/envelope/";>
<soap-env:Body>
<a:put xmlns:a="http://mysite.com/a.xsd";>
<a:A>
<a:A1>
<a:A11>a_string_here</a:A11>
</a:A1>
<a:A2>a_string_here</a:A2>
</a:A>
</a:put>
</soap-env:Body>
</soap-env:Envelope>
i.e, the 'parameters' to the operation 'put' are more kind of a xml structure that could get deeper than for example a:A11. Then the service class' involved method ( for that operation...) would extract the needed xml data to be passed to a back-end system component.
Is this reasoning ok? I mean, choosing to use Axis message style in this case?
My first idea was to develop the service class (and needed handlers) and then get the wsdl via the '?wsdl' trick. That way I suppose users of the service can create clients.
But after reading this conversation I got confused:
- is it better to design the wsdl first? [That way you can follow the steps in your article...]
- isn't designing-the-wsdl-first the only way to get the wrapped convention working?
- or is my first idea ok as well?
What I try is to implement my service(s) the right way from the beginning and using concepts properly.
Thanks in advance. Regards, -R
I apologize -- I'm not an expert on the Axis message handler, so I'm not
100% sure about this answer -- perhaps someone else can pipe up and correct
me if I'm wrong.
But here's my reasoning about how it should work:
When using a message style service with a message signature such as Public void method( SOAPEnvelope req, SOAPEnvelope resp )
you're telling Axis not to process the SOAP body, so it will dispatch the
method before processing the child element of SOAP Body (the operation
name). Option #1 won't work. You have to use option #2: one method for all
operations. Your method must process the XML, extract the operation from the
message, and perform the requested operation.
I'm not quite sure how Axis goes about dispatching the method -- I'm
assuming that the message provider permits only one method to be associated
with a service URL. I'm also not quite sure how to configure this service
such that the WSDL associated with it actually exposes the wrapped
doc/literal interface. But you should be able to do so.
- Anne
-----Original Message-----
From: Rafael Gomez [mailto:[EMAIL PROTECTED] Sent: Monday, September 27, 2004 2:15 PM
To: [EMAIL PROTECTED]
Subject: Re: message style SOAP service
Anne,
I am following this conversation and I would like to ask one more thing based on the wsdl example at http://www.burtongroup.com/weblogs/annethomasmanes/archives/cat_apache_axis.
html
Suppose we add 2 new 'operations' to the service, say 'multiply' and 'divide', both taking again two 'int' arguments and returning an 'int'.
We could complete the wsdl using the same approach used for the operation 'add' (, i.e, via wrapped convention)
Suppose I choose Axis MESSAGE programming style for the service class, and for example, the method signature 'public void method( SOAPEnvelope req, SOAPEnvelope resp )'.
What's the normal/correct thing to do when implementing my service class?:
1) implementing one method per operation, i.e:
public void add(SOAPEnvelope req, SOAPEnvelope resp) {....} public void multiply(SOAPEnvelope req, SOAPEnvelope resp) {....} public void divide(SOAPEnvelope req, SOAPEnvelope resp) {....}
Would Axis call correct method based on the operation name (i.e, based on the wrapper element name in soap request) ?
If yes, when is the correct method determined? Who finds the correct method?
2) implementing only one method for all the operations, i.e:
public void myMethod(SOAPEnvelope req, SOAPEnvelope resp) {....}
Then myMethod could extract 'operation name' and parameters from the request, compute and create/modifiy the resp.
-R
Message examples from a client:
<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope xmlns:xsd="http://www.w3.org/2001/XMLSchema";
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/";
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance";>
<SOAP-ENV:Body>
*<add*>
<arg1 xsi:type="xsd:int">*2*</arg1>
<arg2 xsi:type="xsd:int">*5*</arg2>
*</add*> </SOAP-ENV:Body>
</SOAP-ENV:Envelope>
<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope xmlns:xsd="http://www.w3.org/2001/XMLSchema";
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/";
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance";>
<SOAP-ENV:Body>
*<multiply*>
<arg1 xsi:type="xsd:int">*2*</arg1>
<arg2 xsi:type="xsd:int">*5*</arg2>
*</**multiply*> </SOAP-ENV:Body>
</SOAP-ENV:Envelope>
smime.p7s
Description: S/MIME Cryptographic Signature
