Re: Problems with wsdl2java

Sun, 13 Apr 2008 10:12:09 -0700 

Simon Laws wrote:

On Fri, Apr 11, 2008 at 3:37 PM, Simon Laws <[EMAIL PROTECTED]>
wrote:



On Fri, Apr 11, 2008 at 7:51 AM, Dave Sowerby <[EMAIL PROTECTED]>
wrote:


Hey Simon,

Thanks for the response.

Indeed this is a change in Tuscany behaviour - using the same service
running under 1.0-incubating or 1.1-incubating the WSDL generated is
as expected.  This problem only appears to have started recently with
1.2.

Cheers,

Dave.

On Thu, Apr 10, 2008 at 10:40 PM, Simon Laws <[EMAIL PROTECTED]>
wrote:

On Thu, Apr 10, 2008 at 12:29 PM, Dave Sowerby <[EMAIL PROTECTED]

 wrote:



 > Hi,
 >
 > I'm currently facing issues when attmepting to utilise the wsdl
 > generated by a service exposed using binding.ws, when I use

wsdl2java

 > with this wsdl I get the following exception:
 >
 > IWAB0399E Error in generating Java from WSDL:  java.io.IOException:
 > Emitter failure.  Cannot find endpoint address in port
 > ServiceRequestPortType__SOAPHTTPPort in service
 > ServiceRequestPortType__ServiceLocator
 >    java.io.IOException: Emitter failure.  Cannot find endpoint
 > address in port ServiceRequestPortType__SOAPHTTPPort in service
 > ServiceRequestPortType__ServiceLocator
 >    at
 >

org.apache.axis.wsdl.toJava.JavaServiceImplWriter.writeFileBody(JavaServiceImplWriter.java:189)

 >    at

org.apache.axis.wsdl.toJava.JavaWriter.generate(JavaWriter.java:127)

 >    at
 >

org.apache.axis.wsdl.toJava.JavaServiceWriter.generate(JavaServiceWriter.java:112)

 >    at
 >

org.apache.axis.wsdl.toJava.JavaGeneratorFactory$Writers.generate(JavaGeneratorFactory.java:421)

 >    at org.apache.axis.wsdl.gen.Parser.generate(Parser.java:476)
 >    at org.apache.axis.wsdl.gen.Parser.access$000(Parser.java:45)
 >    at

org.apache.axis.wsdl.gen.Parser$WSDLRunnable.run(Parser.java:362)

 >    at java.lang.Thread.run(Unknown Source)
 >
 > I've diffed a previously functioning wsdl against the currently

(RC3a)

 > generated wsdl file, the difference causing this problem appears to

be

 > the additional lines of:
 >
 >  <wsdl:service name="ServiceRequestPortType__Service">
 >    <wsdl:port name="ServiceRequestPortType__SOAPHTTPPort"
 > binding="ns2:ServiceRequestPortType__SOAPBinding">
 >    </wsdl:port>
 >  </wsdl:service>
 >
 > Which without an address is causing wsdl2java to fail.
 >
 > Has anyone seen this before?  Or does anyone have any suggestions?
 >
 > Cheers,
 >
 > Dave.
 >
 > --
 > Dave Sowerby MEng MBCS
 >
 >

---------------------------------------------------------------------

 > To unsubscribe, e-mail: [EMAIL PROTECTED]
 > For additional commands, e-mail: [EMAIL PROTECTED]
 >
 >
 Hi Dave

 I don't have an immediate suggestion so I'd like to understand if

this is a

 change in behavior in the Tuscany code that you are now seeing. I.e.

The

 previously functioning WSDL that you diffed against. Was that also

generated

 by Tuscany in the past? If so I'll go look at what changed and why.

 As an aside I saw a post from Simon Nash saying that he is looking at

the

 WSDL generation story afresh so hopefully we can make this runtime vs
 development story much more consistent.

 Regards

 Simon




--
Dave Sowerby MEng MBCS

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Dave

As this effect seems to be erroneous in several ways and is different from
1.1 can you raise a high priority JIRA so we can track it.

Thanks

Simon



As a work around while we fix this you could try specifying a WSDL port
manually in the composite. E.g. I tried this on samples/callback-ws-service
by changing the binding.ws as follows.

    <component name="MyServiceComponent">
        <implementation.java class="myserver.MyServiceImpl" />
        <service name="MyService">
            <interface.java interface="myserver.MyService"
                callbackInterface="myserver.MyServiceCallback" />
            <binding.ws uri="http://localhost:8086/MyServiceComponent";
wsdlElement="http://myserver#wsdl.port(MyService/MyServiceSOAP11port_http)"
/>
            <callback>
                <binding.ws />
            </callback>
        </service>
    </component>

I looked in the generated wsdl to work out what the namespace/service/port
should be. The addition of wsdlElement="
http://myserver#wsdl.port(MyService/MyServiceSOAP11port_http)" has the
effect of turning off the extra service/binding that now gets added
otherwise.


This workaround can't be used with calculator-ws-webapp, which
doesn't have any explicit WSDL.  This bug makes the WSDL-less
scenario close to unusable, which is pretty serious.

  Simon


I was a little surprised it allowed me to use both uri= and wsdlElement= but
that's another subject.

Regards

Simon




---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to