RE: binding/operation/infault|outfault?

This thread seems to be languishing, and I don't yet know whether it can
be coalesced into a clear last call comment.  Do you want such a comment
filed?  If so can I characterize it simply as "allow SOAP module on
binding/fault"?

> -----Original Message-----
> From: public-ws-desc-comments-request@w3.org [mailto:public-ws-desc-
> comments-request@w3.org] On Behalf Of Bob Cunnings
> Sent: Thursday, September 09, 2004 10:47 AM
> To: public-ws-desc-comments@w3.org
> Subject: Re: binding/operation/infault|outfault?
> 
> 
> Hello,
> 
> It's not that I want to bind a common fault differently to various
> operations. Rather, I wanted to somehow associate a particular SOAP
module
> with a particular Interface Fault component. It doesn't necessarily
have
> to
> be at the operation level.
> 
> The purpose in mind was to specify a particular soap module (embodied
in a
> header element) that would be present in a particular Fault message to
> carry
> additional information "out of band" with respect to the application
> generating the Fault. This module is not associated with any other
> message.
> 
> IIRC when the binding/operation/outfault was available, the the
> association
> could be made via a wsoap:module component.
> 
> Since faults are now at the interface level, would it make sense for
> binding/fault to contain a SOAP module component? That would do the
job,
> unless I'm missing something here.
> 
> Thanks,
> 
> RC
> 
> 
> I believe they were removed when we lifted faults to the interface
level
> and introduced Interface Fault components. Since faults are now shared
> by multiple operations, they can only be bound once in an interface
> by using a Binding Fault component.
> 
> For example, it's not possible for a single fault, used by three
> different operations, to result in SOAP faults with different fault
> codes depending on which operation was being invoked when the fault
> was raised. I vaguely remember the WG discussing this case at
> the January 2004 F2F hosted by Sonic.
> 
> If there is a genuine need for allowing different operations in the
> same interface to bind the same fault differently, we may want
> to (re-)introduce a Binding Fault Reference component. So please
> tell us more about your use case.
> 
> Thanks,
> Roberto
>

Received on Thursday, 30 September 2004 00:13:50 UTC