Intermediaries, faults, req/resp (was: Edtodo is now uptodate)

So Henrik seems to be saying this is out of scope and should be defined by
the corresponding routing/forwarding feature only, which sounds fine. Do
we want some additional text saying just that or is the spec ok as is?

Jean-Jacques.

Marc Hadley wrote:

> >>> 289: What does an intermediary do when it receives a fault? Is the
> >>> fault guaranteed to be passed on to the original sender (
> >> or previous
> >>> intermediary)?
> >>>
> >>
> >> Since we don't indicate what happens when faults are
> >> generated in the first place, I think the most we can say is
> >> that intermediaries MAY forward fault messages.
> >>
> >> However, I am wondering whether this is not raising a deeper
> >> issue, which is how intermediaries forward response messages.
> >> I think sections 2.7.* are written from the POV of request
> >> messages; do they cover adequately response messages?
> >
> > Good question. What do the other editors think?
> >
> I think this is a deeper issue. Does the behaviour differ depending
> on whether the fault is due to something the intermediary did or a
> prior node in the message path ? Is there a difference between a
> request and response message as far as intermediaries are
> concerned ? How would an intermediary determine the difference ?
>
> Marc.

Henrik wrote:

> >However, I am wondering whether this is not raising a deeper
> >issue, which is how intermediaries forward response messages.
> >I think sections 2.7.* are written from the POV of request
> >messages; do they cover adequately response messages?
>
> I think so: As we don't say anything about how the forwarding feature is
> defined for any SOAP messages other than stating that it is a feature,
> this would also apply to SOAP faults--they are just a certain type of
> SOAP messages.
>