Hopefully implementers aren't confused
by this ... well it confused at least the first implementator in our
org -- because the reality is no one, let alone an implementor that has
been asked to update an implementation based on a section in the spec,
reads the whole/beginning of the spec. Anyway, thanks for the
clarification.
-Mike-
Rich Thompson wrote:
Now I see what you are asking. We
added
the statement you guessed at to v2's Section 2, namely;
"Whenever
this specification uses the prefix "wsrp:" in what appears to
be a QName, it is referring to the URI urn:oasis:names:tc:wsrp:v2:types."
While the use with modes, etc is
literal,
since the underlying type is string rather than QName, the intent is to
point people in the QName direction. Hopefully implementers aren't
confused
by this ...
Rich
What I am asking is whether the name (excluding
namespace)
of the event is "eventHandlingFailed" or "wsrp:eventHandlingFailed"?
The spec seems to suggest the later but I assume we want the former.
Are you saying that elsewhere in the spec we define that references
to wsrp: when used for a value that is a QName implies the wsrp (types)
namespace? I don't recall this and even if we do worry about confusion
with uses that aren't QNames such as modes.
-Mike-
Rich Thompson wrote:
These are defined as elements within our types namespace (i.e.
identical
to the fault messages). The Axis tooling I have been testing with
places
them in the right package/namespace. I'm not sure if you are raising a
tooling issue or if I am just missing the essence of your question.
Rich
Section 6.11 defines two wsrp events:
wsrp:eventHandlingFailed and wsrp:newNavigationalContextScope.
The wsdl defines these as eventHandlingFailed and
newNavigationalContextScope.
I assume the wsdl is the correct way as events are namespace. So
is
this just a simple editorial change?
-Mike-
|