joel-shemtov commented on issue #345: When failing to close conduit, distinct read faults from write faults URL: https://github.com/apache/cxf/pull/345#issuecomment-347911005 Hi Sergey, I checked and found that there actually is a method called testTimeout() in _org.apache.cxf.systest.dispatch.DispatchClientServerTest_. However I'm not sure if it really tests timeout. This is the current code on master ` @Test public void testTimeout() throws Exception { //CXF-2384 URL wsdl = getClass().getResource("/wsdl/hello_world.wsdl"); assertNotNull(wsdl); //pick one of the other service/ports that would have an address //without a service running QName otherServiceName = new QName("http://apache.org/hello_world_soap_http";, "SOAPProviderService"); QName otherPortName = new QName("http://apache.org/hello_world_soap_http";, "SoapProviderPort"); SOAPService service = new SOAPService(wsdl, otherServiceName); assertNotNull(service); Dispatch<SOAPMessage> disp = service .createDispatch(otherPortName, SOAPMessage.class, Service.Mode.MESSAGE); disp.getRequestContext().put(BindingProvider.ENDPOINT_ADDRESS_PROPERTY, "http://localhost:"; + TestUtil.getPortNumber("fake-port") + "/SOAPDispatchService/SoapDispatchPort"); DispatchImpl<?> dispImpl = (DispatchImpl<?>)disp; HTTPConduit cond = (HTTPConduit)dispImpl.getClient().getConduit(); cond.getClient().setConnectionTimeout(500); InputStream is = getClass().getResourceAsStream("resources/GreetMeDocLiteralReq.xml"); SOAPMessage soapReqMsg = MessageFactory.newInstance().createMessage(null, is); assertNotNull(soapReqMsg); try { disp.invoke(soapReqMsg); fail("Should have faulted"); } catch (SOAPFaultException ex) { fail("should not be a SOAPFaultException"); } catch (WebServiceException ex) { //expected assertTrue(ex.getCause().getClass().getName(), ex.getCause() instanceof java.net.ConnectException || ex.getCause() instanceof java.net.SocketTimeoutException); } dispImpl.close(); } ` To me it seems that the usage of "fake-port", guarantees that only ConnectionException gets caught. SocketTimeoutException is never actually thrown. As far as I understand, these are two different scenarios. Shouldn't they be in two separate tests? As you suggested I'll try to run it locally. When I have it modified and tried, I'll open a new pull request.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
