"go away and be happy with what you have"
:-)
Seriously, that's an interesting catch. This syntax basically comes
from IONA's commercial products which have been doing it this way for
YEARS (like 2001 or so). You're the first to catch that it's
invalid. Nice job.
As part of the discussion for 2.2 features, updating the JMS transport
was brought up. One of the things we'd like to do is get it to
implement the proposed spec:
http://www.w3.org/Submission/SOAPJMS/
Dan
On Jul 9, 2008, at 3:45 PM, Dan Retzlaff wrote:
Right, this is still a SOAP binding since the first element of
<wsdl:binding/> is <soap:binding/>, and what goes over the wire is
wrapped
in a SOAP envelope. It's the binding's transport that is JMS.
By analogy, the WSDL spec includes this example for SOAP over SMTP:
<wsdl:binding name="StockQuoteSoap" type="tns:StockQuotePortType">
<soap:binding style="document" transport="http://example.com/smtp
"/>
I can see why CXF was done it this way. The WSDL spec does not allow
the
SOAP binding to have any extensibility elements:
<schema xmlns="http://www.w3.org/2000/10/XMLSchema";
xmlns:soap="http://schemas.xmlsoap.org/wsdl/soap/";
targetNamespace="http://schemas.xmlsoap.org/wsdl/soap/";>
...
<element name="address" type="soap:addressType"/>
<complexType name="addressType">
<attribute name="location" type="uriReference" use="required"/>
</complexType>
And apparently CXF requires additional stuff like "destinationStyle"
and
"jndiConnectionFactoryName". So I guess the question is whether this
non-compliance is intentional. Any insight or guidance, including
"go away
and be happy with what you have" is appreciated. :)
Dan
On Wed, Jul 9, 2008 at 12:11 PM, Glen Mazza <[EMAIL PROTECTED]>
wrote:
Oops--this is SOAP over JMS. Never mind (I think).
Glen Mazza wrote:
I may be wrong here but that's just for the SOAP binding within WSDL
(which has other bindings, namely the HTTP one) A JMS binding with
WSDL
would not be relevant for the SOAP-binding rule below then.
Glen
Dan Retzlaff wrote:
The jms_queue and jms_pubsub samples configure their <wsdl:port/
>s with
a
<jms:address/> element instead of a <soap:address/> element. This
looks
like
the only way to get CXF's JMS transport to actually work, but I
believe
it's
technically invalid. According to the WSDL spec at
http://www.w3.org/TR/wsdl#_soap:address:
3.8 soap:address
The SOAP address binding is used to give a port an address (a
URI). A
port
using the SOAP binding MUST specify exactly one address. The URI
scheme
specified for the address must correspond to the transport
specified by
the
soap:binding.
Is this discrepency worthy of a JIRA report? I'm guessing this
URI-based
transport specification isn't as easy to do with the current
implementation,
but looking through the forum history, I'm not the first to be
confused
by
this. In my case XMLSpy complains every time I try to validate my
CXF-compatible WSDLs.
Regards,
Dan
--
View this message in context:
http://www.nabble.com/Invalid-WSDL-for-SOAP-over-JMS-tp18367273p18368644.html
Sent from the cxf-user mailing list archive at Nabble.com.
---
Daniel Kulp
[EMAIL PROTECTED]
http://www.dankulp.com/blog
