Billy Fence

Hi all,
I have this problem with weblogic 7 clientgen ant task. It is not able to generate client stubs from a wsdl, deployed webservice on a wasp webserver. Axis 1.0 manages fine. It seems that clientgen is not able to generate the client (de)serializers properly. It throws a WSDLParseException and starts nagging about "Cannot have more than one return part. check the parameterOrder specified for the operation: DefaultOperation[ bladiebla"
I think this is weired because there's no problem generating stubs with Axis10.
Is anyone out there using Weblogic 7 clientgen on a wsdl from a different webservice server vendor?
Thank you,
Billy

Hi, trying to create a client call using stubs generated by weblogic (7) tool clientgen on a wasp 4.6 wsdl. I'm stuck with the following error:
java.lang.NoSuchMethodError
at weblogic.xml.schema.binding.internal.TypeMappingEntryImpl.<init>(TypeMappingEntryImpl.java:61)
at weblogic.xml.schema.binding.internal.XSDTypeMapping.addEntries(XSDTypeMapping.java:724)
at weblogic.xml.schema.binding.internal.XSDTypeMapping.createPrimaryMapping(XSDTypeMapping.java:149)
at weblogic.xml.schema.binding.internal.XSDTypeMapping.createInstance(XSDTypeMapping.java:135)
at weblogic.xml.schema.binding.internal.XSDTypeMapping.<clinit>(XSDTypeMapping.java:116)
at weblogic.xml.schema.binding.internal.TypeMappingFactoryBase.createDefaultMapping(TypeMappingFactoryBase.java:18)
at weblogic.webservice.core.encoding.DefaultRegistry.<init>(DefaultRegistry.java:77)
at weblogic.webservice.core.rpc.ServiceImpl.<init>(ServiceImpl.java:68)
at weblogic.webservice.core.rpc.ServiceImpl.<init>(ServiceImpl.java:57)
at weblogic.webservice.core.rpc.ServiceFactoryImpl.createService(ServiceFactoryImpl.java:36)
Anyone any ideas? It looks like I'm missing a jar in my classpath but which? I'm using the ones recommended by weblogic to make client calls.
Thanks for the help.
B

Got it sorted. Thanks for the help.
B.

Alternatively I have an xmap file, which should be availlable on the client (?). Can anyone tell me where a client should put that file or reference that file in the code or ....?
Cheers,
Willem

Thanks for your explaination. I assumed the generated wsdl would contain that kind of information, shouldn't it?

Lasse, your explaination works in an ideal world. I'm starting to look into webservices so I can't speak too much but having heard from experts in our place there's a lot not right because of soap 1.1. I've seen wsdl's that work fine with wasp 4.5 or 4.6 and weblogic 6.1 and .net but when you use the same wsdl and weblogic 7, axis etc. etc. it all doesn't work anymore. Client calls simply do not what you'd expect. So in theory it works. In practice with a lot of vendors and a not fully matured soap definition, as a developer, you're bound to find trouble.

The thing is, I think I'm doing it right. I specified the wsdl location and the service name from my wsdl. Sofar it works, but then when I invoke my Call instance I get this exception:
java.rmi.RemoteException: java.beans.IntrospectionException: Due to undefined mapping between java.lang.Object.class and XML Schema WASP can not send instance of java.lang.Object.; nested exception is:
org.idoox.xmlrpc.MessageProcessingException: java.beans.IntrospectionException: Due to undefined mapping between java.lang.Object.class and XML Schema WASP can not send instance of java.lang.Object.
org.idoox.xmlrpc.MessageProcessingException: java.beans.IntrospectionException: Due to undefined mapping between java.lang.Object.class and XML Schema WASP can not send instance of java.lang.Object.
at com.idoox.wasp.serialization.xml.XMLReflectionDeserializer.deserialize(XMLReflectionDeserializer.java:100)
at com.idoox.wasp.serialization.xml.XMLBasicConstructDeserializer.deserialize(XMLBasicConstructDeserializer.java:126)
at com.idoox.wasp.serialization.xml.XMLSimpleDeserializer.deserialize(XMLSimpleDeserializer.java:138)
at com.idoox.wasp.serialization.SerializationHelper.deserializeElement(SerializationHelper.java:349)
at com.idoox.wasp.serialization.WaspSerializationHelper.deserialize(WaspSerializationHelper.java:174)
at com.systinet.wasp.rpc.WaspCallImpl.deserializeFromXml(WaspCallImpl.java:1388)
at com.systinet.wasp.rpc.WaspCallImpl.readResponse(WaspCallImpl.java:1094)
at com.systinet.wasp.rpc.WaspCallImpl.deserializeResponse(WaspCallImpl.java:955)
at com.systinet.wasp.rpc.WaspCallImpl.invoke(WaspCallImpl.java:491)
I'm using wasp 4.6

Anyone any idea?
Thx!
Billy

Bearing in mind the wsdl.

To create a Service I use a ServiceFactory, then I create a call, specify the operation and the arguments before invoking the call. Can someone in a clear way explain to me what the arguments for these methods should point to? Cheers.

Don't really get your explaination. Maybe I'm trying to say can you create more more than one nodemanager per physical machine. As far as I see it I'm trying to create a managed server on the same server as the administration server. When I try to start this managed server I invoke a nodemanager which is located on the administration server, and I wonder if that is possible?

Hi,
Really hope someone out there can help me. I'm using standard weblogic 6.1 install with only one administration server 'myserver'. By default this is tied to machine (none), listens to 7001 and has an empty listen address. When I want to create a new logical server, for example one named applicserver I need to bind that to a machine. Since the machine listbox only shows me (none) I try that and start to run it. This is what happens:
<08-Aug-03 15:24:29 CEST> <Error> <NodeManager> <Could not start server 'wasp-adminserver' via Node Manager - reason: '[SecureCommandInvoker: Could not create a socket to the NodeManager running on host 'localhost:5555' to execute command 'online applicserver', reason: Connection refused: connect. Ensure that the NodeManager on host 'localhost' is configured to listen on port '5555' and that it is actively listening]'>

So I thought let's create a new machine and call that localhost that listens to another port, for example 7999. Then tried to bind the applicserver to that new machine and tried to start it. Again the above output. It drives me nutts. I don't have a helping hand close by and the weblogic documentation is too much to go through, so I'm stuck. Can anyone out there explain what I'm doing wrong.
Cheers
Billy

I will repost it in the bea/weblogic forum.
Kiitos
p.s. If we all would look at the documentation there would be no forum ;-)
I'm really stuck with this one.....

Is there anyone out there who can tell me how to correctly create and deploy a wasp.war servlet on a weblogic 6.1 standalone server domain. I just can't get it done.
I'm using wasp 4.6 J2eeIntegrate --installation-type=weblogic --htpp-port=??? -https-port=??? --shared
I presume I should use 7001 7002 and deploy the servlet and call http://localhost:???/<context>/wasp/admin/console but I just can't seem to get it to work.
Thx.

Hi,
Really hope someone out there can help me. I'm using standard weblogic 6.1 install with only one administration server 'myserver'. By default this is tied to machine (none), listens to 7001 and has an empty listen address. When I want to create a new logical server, for example one named applicserver I need to bind that to a machine. Since the machine listbox only shows me (none) I try that and start to run it. This is what happens:

<08-Aug-03 15:24:29 CEST> <Error> <NodeManager> <Could not start server 'wasp-adminserver' via Node Manager - reason: '[SecureCommandInvoker: Could not create a socket to the NodeManager running on host 'localhost:5555' to execute command 'online applicserver', reason: Connection refused: connect. Ensure that the NodeManager on host 'localhost' is configured to listen on port '5555' and that it is actively listening]'>

So I thought let's create a new machine and call that localhost that listens to another port, for example 7999. Then tried to bind the applicserver to that new machine and tried to start it. Again the above output. It drives me nutts. I don't have a helping hand close by and the weblogic documentation is too much to go through, so I'm stuck. Can anyone out there explain what I'm doing wrong.
Cheers
Billy

More times than not, performance tuning is done prematurely, which usually results in tuning code that may not make a difference in the end or that is already optimized behind the scenes by the compiler or JIT, instead of identifying the actual bottleneck in the system using a profiler.

That's what I did. It's loosing to much time in the client side where processing of the errors is done.

I used a profiler, found the bottlenecks. Working on each of them seperatly. I hear a lot of people say concentrate on the real bottlenecks. Sure, but usually re-design of the architecture and refactoring of some code at some stage will sort that out, if you have the time and resources for that. Still I think that when you program you should be aware of performance all the time. I see a lot of programmers use '+' with Strings whilst all books say use StringBuffer. Ok, it might be a small thing but if you're a real-time systems programmer they'll hang you by the goolies with an attitude like that. What I'm trying to say is, if there is a better way, get to know it, use it and don't always hide behind the re-design dogma. You guys know as well as I that projects hardly ever go the way you want and there might never be the time to sort out the big issues which cost a lot of time (=$$$).
Anyway thanks for the replies.