cxf-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Glynn, Eoghan" <eoghan.gl...@iona.com>
Subject RE: [jira] Closed: (CXF-36) Fault Handling
Date Wed, 27 Sep 2006 10:16:25 GMT

Hi Dan,

I came across a similar issue at the weekend when system testing
wsa:faultTo support (I've left the test case disabled out for the
moment).

A quick look into the issue suggested that the problem lay in the
Soap11FaultInterceptor only being added to the in-fault-chain as opposed
to the normal in-chain, so that when the response is dispatched to the
in-chain by ClientImpl.onMessage() there's no interceptor responsible
for mapping from the <soap:Fault> element to the correct exception type.

The questions that occurred to me at the time were:

1. How can ClientImpl.onMessage() choose between the in-chain and
in-fault-chain, i.e. distinguish a priori between an incoming fault and
normal response *before* the message is unmarshalled?

2. Or is the model that the incoming message is always dispatched
initially on the in-chain, and then jumps to the fault chain once its
clear that the payload contains a fault?

3. What's the advantage of having separate normal and fault interceptor
chains when each interceptor has separate entry points for normal and
fault messages (i.e. handleMessage() and handleFault()). Rather than
dispatching ab initio on a separate fault chain, or jumping from the
normal to the fault chain, it would seem simpler to just have a single
chain and distinguish faults via the interceptor entry point. It would
also be more consistent with the JAX-WS Handler model (Why's that
important? Well its likely IMO that some developers will be writing both
JAX-WS Handlers and native CXF Interceptors, so it behooves us not to
introduce gratuitous inconsistencies between these dispatch models
without good reason)

Cheers,
Eoghan

> -----Original Message-----
> From: Dan Diephouse [mailto:dan@envoisolutions.com] 
> Sent: 27 September 2006 04:35
> To: cxf-dev@incubator.apache.org
> Subject: Re: [jira] Closed: (CXF-36) Fault Handling
> 
> Oops, forgot about the client side. I just reopened the 
> issue. I'll look into it later this week.
> 
> - Dan
> 
> Jiang Ning wrote:
> 
> > Hi Dan,
> >
> > I had tried the lastest kit  helloworld demo and tried to run the 
> > fault test in the systest, the fault isn't work yet (the 
> client side 
> > code can't get the Java Exception), can you look into it.
> >
> > Thanks,
> > Willem.
> >
> > Dan Diephouse (JIRA) wrote:
> >
> >>     [ http://issues.apache.org/jira/browse/CXF-36?page=all ]
> >>
> >> Dan Diephouse closed CXF-36.
> >> ----------------------------
> >>
> >>    Resolution: Fixed
> >>
> >> This has been implemented for SOAP 1.1 in current SVN.
> >>  
> >>
> >>> Fault Handling
> >>> --------------
> >>>
> >>>                Key: CXF-36
> >>>                URL: http://issues.apache.org/jira/browse/CXF-36
> >>>            Project: CeltiXfire
> >>>         Issue Type: Task
> >>>           Reporter: Dan Diephouse
> >>>        Assigned To: Dan Diephouse
> >>>            Fix For: 2.0-M1
> >>>
> >>>
> >>> Fault handling isn't quite finished in SVN. Specifically 
> I need to 
> >>> actually wire in the Soap fault handlers and add 
> exception converters.
> >>>   
> >>
> >>
> >>  
> >>
> >
> 
> 
> --
> Dan Diephouse
> (616) 971-2053
> Envoi Solutions LLC
> http://netzooid.com
> 
> 
> 

Mime
View raw message