You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@cxf.apache.org by Benson Margulies <bi...@basistech.com> on 2007/11/01 21:21:18 UTC

Life as a code generator: questions of style

The JavaScript generator walks into a bar ...

Well, really, it walks up to ServiceInfo. It examines the bindings. It
ignores all bindings except SoapBindingInfo bindings, since no way is
someone going to try to use JBI to connect a web browser to a web
service.

The SoapBinding says 'document' or 'rpc'. In fact, it says this one
operation at a time, due to the SoapOperationInfo.

That leaves the question, in the document case, of whether we're wrapped
or literal.

Now, the ReflectionServiceFactoryBean negotiates with the service
configuration to decide whether to wrap things up or not. The question
is, does it leave behind a flag telling a consumer of the ServiceInfo
(and the rest of the model) what it did? Well, if an operation is
wrapped, its operation gets an UnwrappedOperation. So if it's not
wrapped, the slot for the unwrapped operation is null? 

OperationInfo.isUnwrappedCapable then seems to indicate that the
operation is wrapped, but the name of this predicate worries me. If it
returns true, should the JavaScript code generator just generate a
'wrapped' style of API? That is what is happening in
org.apache.cxf.tools.wsdlto.frontend.jaxws.processor.internal.mapper.Met
hodMapper, if I'm reading it correctly. There seems to be a comment
problem in there, btw.