com/sun/ts/tests/jaxws/api/javax_xml_ws/AsyncHandler/Client.java#HandleResponseXMLTest_from_wsappclient
com/sun/ts/tests/jaxws/api/javax_xml_ws/AsyncHandler/Client.java#HandleResponseXMLTest_from_wsappclient_reverse
com/sun/ts/tests/jaxws/api/javax_xml_ws/AsyncHandler/Client.java#HandleResponseXMLTest_from_wsejb
com/sun/ts/tests/jaxws/api/javax_xml_ws/AsyncHandler/Client.java#HandleResponseXMLTest_from_wsservlet

I have questions about TCK jaxws. As I said on the meeting there is bug in TCK's test code (ugly code, see comments below). If I will make small modification to this code, test passes.

Code:
package - com/sun/ts/tests/jaxws/api/javax_xml_ws/AsyncHandler
...
XMLAsyncHandler xah = new XMLAsyncHandler();
while (!xah.isDataReady()); <-- !!!!!
...

public class XMLAsyncHandler implements AsyncHandler<Source> {
...
private boolean dataReady = false;
...
public Source getData()

{ return theData; }

public void handleResponse(Response<Source> res)

{ ... dataReady = true; ... }

}

Instance of the XMLAsyncHandler is registered as ResponseHandler, the second thread cannot see modification in the dataReady prop.
If I make small modification like this
...
while (!xah.isDataReady()) {
try

{ Thread.sleep(50); }

catch (Exception e) {
}
}
...
It works.

Question:
Test passes on AS 5.1.GA with JDK 1.6 U11
http://hudson.qa.jboss.com/hudson/view/TCK5-EAP5-JDK6/job/tck5-eap5-jdk6-jaxws/11/
Test fails on EAP5 CR1, CR2 and CR3 with JDK 1.6 U15
http://hudson.qa.jboss.com/hudson/view/TCK5-EAP5-JDK6/job/tck5-eap5-jdk6-jaxws/25/

Is this introduced by a change in EAP or can we challenge Sun about the test as mentioned above?