Ok thanks, i'll go explore that for a little while...probably I'll be having
more questions while doing that :)

   ...ant

On Dec 14, 2007 7:20 PM, Raymond Feng <[EMAIL PROTECTED]> wrote:

> Yes, creating the WrapperInfo manually is an alternative to generating a
> wrapper style WSDL (in which case the WSDL interface introspection
> produces
> the WrapperInfo).
>
> Thanks,
> Raymond
>
> ----- Original Message -----
> From: "ant elder" <[EMAIL PROTECTED]>
> To: "tuscany-dev" <[email protected]>
> Sent: Friday, December 14, 2007 11:09 AM
> Subject: Re: How to set a specific data binding to be used by a binding?
>
>
> > Ok, so from "unless the binding interface contract is a wrapper style
> WSDL
> > portType" does that mean it really must have a WSDL portType and thats
> > baked
> > deep into the design or is it more "unless the binding interface
> contract
> > is
> > a wrapper style" and it would be possible to just set that up with a
> Java
> > interface? Something like setting a WrapperInfo with
> > o.a.t.s.interfacedef.Operation.setWrapper? Which I guess is assuming its
> > straightforward to create a WrapperInfo object?
> >
> >   ...ant
> >
> > On Dec 14, 2007 5:39 PM, Raymond Feng <[EMAIL PROTECTED]> wrote:
> >
> >>  Hi,
> >>
> >> I just use JMS as an example for bindings that are not naturally
> >> RPC-style.
> >>
> >> Using AXIOM as the databinding will only give you an array of
> OMElements
> >> representing the parameters unless the binding interface contract is a
> >> wrapper style WSDL portType (then you'll receive the wrapper
> OMElement).
> >> We
> >> can enhance the XMLStringDataBinding to handle wrappers but we still
> need
> >> to
> >> express that the binding interface contract is wrapper style in the
> first
> >> place.
> >>
> >> Thanks,
> >> Raymond
> >>
> >> ----- Original Message -----
> >> *From:* ant elder <[EMAIL PROTECTED]>
> >> *To:* Raymond Feng <[EMAIL PROTECTED]>
> >> *Cc:* [email protected]
> >> *Sent:* Friday, December 14, 2007 9:12 AM
> >> *Subject:* Re: How to set a specific data binding to be used by a
> >> binding?
> >>
> >> I'm completely missing the point sorry and I'm not sure I understand
> why
> >> JMS comes into this? I've a binding and I want the wrapped XML. I want
> it
> >> as
> >> a String but I could use DOM or OMElement or what ever else. Right now
> I
> >> (think?) it would work if I just use the Axiom data binding which gives
> >> the
> >> wrapped style of XML and then I'd receive OMElements as the payload
> which
> >> I
> >> could call toString on to get a String, but I'd prefer not to have the
> >> dependency on Axiom.
> >>
> >> I still don't understand why there couldn't be a version of
> >> XMLStringDataBinding that returns wrapped style XML?
> >>
> >>    ...ant
> >>
> >> On Dec 14, 2007 4:55 PM, Raymond Feng < [EMAIL PROTECTED]> wrote:
> >>
> >> > There are two independent factors in the operation-level data
> >> > transformation
> >> > (i.e., converting the payload from one structure to another). In the
> >> > JMSBinding case, if a java interface is used, it's from Object[] to
> >> > JMSMessage (or the text for JMS TextMessage).
> >> >
> >> > 1) The data conversion (from example, a Customer SDO to String)
> >> > 2) The wrapping/unwrapping process. Wrapper style is just a special
> >> > pattern
> >> > for PRC-style payload to messge-style. Other mapping can be used too,
> >> > for
> >> > example, you can create a JMS MapMessage and set the arguments into
> the
> >> > Map.
> >> >
> >> > Since there are two separate concerns, we don't want to have
> >> > XMLStringDataBinding to support this combination only.
> >> >
> >> > Looking at operation-level, we have two high-level transformers
> >> > (idl:input
> >> > <--> idl:input and idl:output <--> idl:output) to deal with payload
> >> > transformation. The transformers handle wrapping/unwrapping and then
> >> > delegate to parameter-level transformers. If the JMS binding wants to
> >> > control how to package the input/output payload, you can register
> >> > special
> >> > transformers  (idl:iput --> jms:input, jms:input --> idl:input,
> >> >  idl:output --> jms:output and jms:output --> idl:ouput). Then in
> your
> >> > transformers, you can wrap/unwrap the XML elements.
> >> >
> >> > Thanks,
> >> > Raymond
> >> >
> >> > ----- Original Message -----
> >> > From: "ant elder" <[EMAIL PROTECTED]>
> >> > To: < [email protected]>
> >> > Sent: Friday, December 14, 2007 3:01 AM
> >> > Subject: Re: How to set a specific data binding to be used by a
> >> > binding?
> >> >
> >> >
> >> >  > On Dec 13, 2007 10:36 PM, ant elder < [EMAIL PROTECTED]> wrote:
> >> > >
> >> > >>
> >> > >>
> >> > >> On Dec 13, 2007 7:22 PM, Raymond Feng <[EMAIL PROTECTED]> wrote:
> >> > >>
> >> > >> <snip>
> >> > >>
> >> > >> Anthoer way you could do is to set the databinding to DOM or
> AXIOM.
> >> > The
> >> > >> > databinding interceptor will pass you an array of elements
> >> > >> > corresponding
> >> > >> > to
> >> > >> > the java arguments.  You can then create the wrapper from the
> >> > >> > child
> >> > >> > args
> >> > >> >
> >> > >> > when it reaches the JMS binding code.
> >> > >> >
> >> > >>
> >> > >> If its possible to fix this like that (by manually adding a
> wrapper
> >> > >> element) then would another way be to have a new data binding that
> >> > does
> >> > >> this
> >> > >> and use that in the binding. Something like a
> >> > >> WrappedXMLStringDataBinding.NAME which is just like the
> >> > >> XMLStringDataBinding.NAME but is implemented to use DOM and
> manually
> >> > >> create a wrapper around the child args? (asking to confirm I'm
> >> > >> understanding
> >> > >> what you're suggesting) If thats possible it seems like it would
> be
> >> > >> a
> >> > >> cleaner approach for binding or implementation type writers.
> >> > >>
> >> > >> If we can do that i'd also use it for implementation.script which
> >> > needs
> >> > >> similar functionality.
> >> > >>
> >> > >>    ...ant
> >> > >>
> >> > >>
> >> > > No replies yet so I had a quick look at doing this...
> >> > >
> >> > > Ideally the WrappedXMLStringDataBinding would be just the same as
> the
> >> > > existing XMLStringDataBinding but that doesn't look like it would
> >> > > work
> >> > -
> >> > > is
> >> > > it possible to have multiple databindings coexisting using the same
> >> > > physical
> >> > > type?
> >> > >
> >> > > So a question from earlier again: Why is XMLStringDataBinding
> >> > producing
> >> > > unwrapped XML? All the uses for this that I have would use wrapped
> >> > style
> >> > > xml
> >> > > so having XMLStringDataBinding produce that would be more
> convenient.
> >> > >
> >> > >   ...ant
> >> > >
> >> >
> >> >
> >>
> >
>
>

Reply via email to