Exception is thrown only the first time an error occur while trying to publish
a webservice with HTTP transport
---------------------------------------------------------------------------------------------------------------
Key: CXF-1589
URL: https://issues.apache.org/jira/browse/CXF-1589
Project: CXF
Issue Type: Bug
Components: Transports
Affects Versions: 2.0.6
Reporter: Kjell Winblad
Fix For: 2.0.7
When a webservice with HTTP defined as transport is published with the
following code and the service is already started at the defined publish
address an exception is thrown which is a correct behaviour. The problem is
that if the same publish code is executed again no error is thrown.
try{
Endpoint end = Endpoint.create(impl);
endpoint = (EndpointImpl) end;
endpoint.setWsdlLocation(wsdlfile);
endpoint.publish();
}catch(WebServiceException e){
System.out.println(e);
//The following error message will be printed if the address is already in
use:
//javax.xml.ws.WebServiceException: org.apache.cxf.interceptor.Fault:
START_UP_SERVER_FAILED_MSG
//org.apache.cxf.interceptor.Fault: START_UP_SERVER_FAILED_MSG
//java.net.BindException: Address already in use
}
try{
Endpoint end = Endpoint.create(impl);
endpoint = (EndpointImpl) end;
endpoint.setWsdlLocation(wsdlfile);
endpoint.publish();
}catch(WebServiceException e){
System.out.println(e);
//No exception is thrown this time even if the publish address is already in
use
}
The problem seems to be in the method
org.apache.cxf.transport.http_jetty.JettyHTTPServerEngine#addServant. The
method has the following structure:
public synchronized void addServant(URL url, JettyHTTPHandler handler) {
if (server == null) {
.... code to start server
} catch (Exception e) {
LOG.log(Level.SEVERE, "START_UP_SERVER_FAILED_MSG", new
Object[] {e.getMessage()});
//problem starting server
try {
server.stop();
server.destroy();
} catch (Exception ex) {
//ignore - probably wasn't fully started anyway
} // Add the following to fix the problem
//finally{
// server = null;
//}
throw new Fault(new Message("START_UP_SERVER_FAILED_MSG", LOG,
e.getMessage()), e);
}
}
The problem is that server field is not null the second time publish is called.
Adding the expression "server = null;" in a finally block for the server.stop()
and server.destroy() try seems to fix the problem.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
