I think the best thing is to refer you to the spec at http://www.w3.org/TR/SOAP/. This shows many valid messages, and the section on SOAP Encoding gets into how to specify data types for parameters.
I believe that your messages are valid. If you supply WSDL for your SOAP services, those messages can probably interoperate quite well with other SOAP implementations that use WSDL, such as Axis and Microsoft (.NET and toolkit). If it is your objective to interoperate as best possible with as many implementations as possible, you will want to emit xsi:type attributes for each data element (and presumably be able to consume these intelligently). Of course, you have to weigh this objective against the resources (your time) you will expend on the implementation. My first experience with SOAP was writing a partial implementation of the 0.9 spec. When the time came to support the remainder of that spec and items added or changed for 1.0, as well as interoperate with other implementations, I chose to migrate to Apache SOAP. Fortunately, I was not constrained against using a servlet container on the server side. Scott Nichol ----- Original Message ----- From: "Bhushan Khanal" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Monday, July 22, 2002 2:29 PM Subject: RE: More on QName problem > Just to give you some background, I using my own soap implementation to > create the message. I trying to write a non servlet container to implement > SOAP services and I was creating mock messages that represented incoming > requests. In addition, I am not using Tomcat or Axis(have my own http and > tcp implementation), so I am quite the Black Sheep of this family. Therefore > its possible that the message could be non confirming to SOAP standards. > Here is my earlier posting that contains the incoming message > http://marc.theaimsgroup.com/?l=soap-dev&m=102704360007200&w=2. > > I was qualifying my parameters with the name space, that was the reason I > ran into this problem in the first place. Do you mean that there should be > an additional attribute that specifies the type? Does that also mean, all > the type information needs to be transmitted with the SOAP message? Could > you point me to an example of a valid SOAP message, that would be very > helpful. > > Thanks, > Bhushan > > > -----Original Message----- > From: Scott Nichol [mailto:[EMAIL PROTECTED]] > Sent: Saturday, July 20, 2002 11:25 PM > To: [EMAIL PROTECTED] > Subject: Re: More on QName problem > > > I would guess it has not been a problem because "name" is only used in the > interop extension branch. Specifically, for parameters with xsi:type > specified or with an href attribute, the code does not rely on "name" in any > way. Most SOAP implementations provide xsi:type for parameters. The most > notable exception is Microsoft, which chose to rely on WSDL for typing. The > "name" variable is used in the branch of code added to attempt to interop > with implementations from Microsoft (and others?) that do not provide > xsi:type. I presume that Apache SOAP users have only exercised this code to > interop with Microsoft, which does not qualify parameters with namespaces. > > I agree that this code should use the local name. I am curious, however, > what SOAP implementation you are using that has namespace qualifiers on the > parameter elements. > > Scott Nichol > > ----- Original Message ----- > From: "Bhushan Khanal" <[EMAIL PROTECTED]> > To: <[EMAIL PROTECTED]> > Sent: Friday, July 19, 2002 6:46 PM > Subject: More on QName problem > > > > Hello again, > > > > Here are some more information that I found on the problem that I have > been > > having with QName and their local names. Please look at the following > > function inside class ParameterSerializer. > > > > public Bean unmarshall(String inScopeEncStyle, QName elementType, Node > src, > > XMLJavaMappingRegistry xjmr, SOAPContext ctx) > > throws IllegalArgumentException > > { > > Element paramEl = (Element)src; > > String name = paramEl.getTagName(); <<---- > > Bean bean = null; > > . > > . > > . > > } > > > > The method getTagName() returns the name of the tag with the name space > > alias. Shouldn't you be extracting the name using getLocalName() instead? > I > > am surprised that this has not been a problem for others. Am I missing > > anything? > > > > Happy Friday! > > > > Bhushan Khanal > > WRQ, Inc. > > 206.217.7636 > > [EMAIL PROTECTED] > > > > WRQ has 21 years of experience providing integration software and services > > for host-intensive environments. To learn more about our Reflection and > > Verastream products, visit http://www.wrq.com/products/. > > > > > > > > > > -- > > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> > > > > > > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> > > -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
