Mailing-List: contact cxf-issues-help@incubator.apache.org; run by ezmlm
Precedence: bulk
Reply-To: cxf-dev@incubator.apache.org
Message-ID: <20617585.1178135295506.JavaMail.jira@brutus>
Date: Wed, 2 May 2007 12:48:15 -0700 (PDT)
From: "Steven E. Harris (JIRA)" <jira@apache.org>
To: cxf-issues@incubator.apache.org
Subject: [jira] Created: (CXF-627) WSDL2Java JAXWS frontend does not handle
 enableAsyncMapping for non-Java-named operations
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

WSDL2Java JAXWS frontend does not handle enableAsyncMapping for non-Java-named operations
-----------------------------------------------------------------------------------------

                 Key: CXF-627
                 URL: https://issues.apache.org/jira/browse/CXF-627
             Project: CXF
          Issue Type: Bug
          Components: Tooling
    Affects Versions: 2.0-RC
         Environment: NA
            Reporter: Steven E. Harris


Per the discussing on the cxf-user mailing list in the thread entitled

  JAX-WS enableAsyncMapping directive not respected
  http://www.nabble.com/Re%3A-JAX-WS-enableAsyncMapping-directive-not-respected-p10292965.html

the class org.apache.cxf.tools.wsdlto.frontend.jaxws.processor.internal.OperationProcessor uses a method called isAddedAsyMethod() to determine whether it has already defined an asynchronous method for a given operation name. Its detection heuristic fails for operation names that don't conform to the Java naming conventions.

For example, an operation name "some-op" will get munged into the Java name "someOp". The isAddedAsyMethod() heuristic looks to see if there are two methods in an interface that share the same operation name, but differ from the operation name in method name -- when dropped to lowercase. My example here meets that criteria, but does not constitute an "extra" asynchronous method being present. It's just tripping on the mistaken assumption that an operation name and method name will be the same save for case.

We need a better collision detection heuristic here -- perhaps one that takes the operation name-to-Java name munging into account.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.