[weld-dev] CDI 2.0 TCK ObserverMethodWithoutNotifyMethodTest failure - ShouldThrow...Definition or Deployment Exception? - spec inconsistency??

Fri Apr 27 02:26:43 EDT 2018

Hi Gordon,

running the test on WFLY (or actually without EE container) gives me the same stacktrace, however the test passes for me.
What is it you fail on exactly?

Debugging a bit shows that @ShouldThrowException annotation (which is guarding the exception in the test) goes through a whole chain of exceptions - having it wrapped should still pass the test.
E.g. it looks at the cause, see if that's assignable, and if it isn't then the method is recursively invoked for Exception.getCause().
Here is a link to Arq. code doing that - https://github.com/arquillian/arquillian-core/blob/master/container/impl-base/src/main/java/org/jboss/arquillian/container/impl/client/container/DeploymentExceptionHandler.java#L63-L66

As for spec point of view...
Wrapping those exceptions meets both criteria mentioned in spec as long as you can get to both of them.
Also, having DefinitionException on top makes sense from user point of view (at least in my opinion) because is it ultimately developer's mistake that he/she did not add the notify() method.
And that subsequently leads to DeploymentException.
Does it make sense?

I'd like to help more but I am not sure whether you are battling some actual test failure or the somewhat clashing sentences in spec itself?

Matej

----- Original Message -----
> From: "Gordon Hutchison" <gordon.hutchison at gmail.com>
> To: weld-dev at lists.jboss.org
> Sent: Thursday, April 26, 2018 5:17:55 PM
> Subject: [weld-dev] CDI 2.0 TCK ObserverMethodWithoutNotifyMethodTest failure - ShouldThrow...Definition or
> Deployment Exception? - spec inconsistency??
> 
> Hi Folks,
> 
> I work in the IBM CDI implementation team.
> 
> We are running the CDI 2.0 TCK/CDS and get a test failure on:
> 
> org.jboss.cdi.tck.tests.extensions.lifecycle.broken.observerMethod.ObserverMethodWithoutNotifyMethodTest
> 
> that I need some advice on.
> 
> The test has:
> 
> import javax.enterprise.inject.spi.De ploymentException;
> 
> @ShouldThrowException(DeploymentException.class)
> 
> 
> // unable to process ObserverMethodConfigurator due to missing notify method
> impl
> @Test
> @SpecAssertions({ @SpecAssertion(section = AFTER_BEAN_DISCOVERY, id = "ee")
> })
> public void unableToProcessObserverMethodConfigurator() {
> }
> 
> This behaviour is specified in
> https://docs.jboss.org/cdi/spec/2.0/cdi-spec-with-assertions.html#after_bean_discovery
> (search for "ee)" )
> 
> ee) If the container is unable to process the ObserverMethodConfigurator it
> automatically detects the problem and treats it as a deployment problem.
> ObserverMethodWithoutNotifyMethodTest.unableToProcessObserverMethodConfigurator()
> 
> 
> However what we see coming back from Weld is not a javax DeploymentException
> but a javax DefinitionException wrapping a weld DeploymentException cause
> 
> Stack Dump = javax.enterprise.inject.spi.De finitionException:
> org.jboss.weld.exceptions.DeploymentException: WELD-000179:
> ObserverMethodConfigurator created by
> org.jboss.cdi.tck.tests.extensions.lifecycle.broken.observerMethod.AfterBeanDiscoveryObserver at d31423c
> cannot be processed
> at
> org.jboss.weld.bootstrap.events.AfterBeanDiscoveryImpl.finish(AfterBeanDiscoveryImpl.java:180)
> <<<<<<< Problem????
> at
> org.jboss.weld.bootstrap.events.AfterBeanDiscoveryImpl.fire(AfterBeanDiscoveryImpl.java:76)
> at org.jboss.weld.bootstrap.WeldStartup.deployBeans(WeldStartup.java:446)
> at org.jboss.weld.bootstrap.WeldBootstrap.deployBeans(WeldBootstrap.java:83)
> at
> com.ibm.ws.cdi.impl.CDIContainerImpl.startInitialization(CDIContainerImpl.java:149)
> at
> com.ibm.ws.cdi.liberty.CDIRuntimeImpl.applicationStarting(CDIRuntimeImpl.java:400)
> at
> com.ibm.ws.container.service.state.internal.ApplicationStateManager.fireStarting(ApplicationStateManager.java:28)
> at
> com.ibm.ws.container.service.state.internal.StateChangeServiceImpl.fireApplicationStarting(StateChangeServiceImpl.java:50)
> at
> com.ibm.ws.app.manager.module.internal.DeployedAppInfoBase.preDeployApp(DeployedAppInfoBase.java:383)
> at
> com.ibm.ws.app.manager.module.internal.DeployedAppInfoBase.deployApp(DeployedAppInfoBase.java:412)
> at com.ibm.ws.app.manager.war.int
> ernal.WARApplicationHandlerImpl.install(WARApplicationHandlerImpl.java:65)
> at
> com.ibm.ws.app.manager.internal.statemachine.StartAction.execute(StartAction.java:140)
> at
> com.ibm.ws.app.manager.internal.statemachine.ApplicationStateMachineImpl.enterState(ApplicationStateMachineImpl.java:1258)
> at
> com.ibm.ws.app.manager.internal.statemachine.ApplicationStateMachineImpl.run(ApplicationStateMachineImpl.java:873)
> at
> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
> at
> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
> at java.lang.Thread.run(Thread.java:748)
> Caused by: org.jboss.weld.exceptions.DeploymentException: WELD-000179:
> ObserverMethodConfigurator created by
> org.jboss.cdi.tck.tests.extensions.lifecycle.broken.observerMethod.AfterBeanDiscoveryObserver at d31423c
> cannot be processed
> at
> org.jboss.weld.bootstrap.events.AfterBeanDiscoveryImpl$ObserverRegistration.initObserverMethod(AfterBeanDiscoveryImpl.java:368)
> at
> org.jboss.weld.bootstrap.events.AfterBeanDiscoveryImpl.processObserverRegistration(AfterBeanDiscoveryImpl.java:268)
> at
> org.jboss.weld.bootstrap.events.AfterBeanDiscoveryImpl.finish(AfterBeanDiscoveryImpl.java:177)
> ... 16 more
> Caused by: org.jboss.weld.exceptions.DefinitionException: WELD-000415: Custom
> implementation of observer method does not override either notify(T) or
> notify(EventContext<T>):
> org.jboss.weld.bootstrap.events.configurator.ObserverMethodConfiguratorImpl at 7e164c01
> at
> org.jboss.weld.bootstrap.events.configurator.ObserverMethodConfiguratorImpl$ImmutableObserverMethod.<init>(ObserverMethodConfiguratorImpl.java:272)
> at
> org.jboss.weld.bootstrap.events.configurator.ObserverMethodConfiguratorImpl.complete(ObserverMethodConfiguratorImpl.java:234)
> at
> org.jboss.weld.bootstrap.events.AfterBeanDiscoveryImpl$ObserverRegistration.initObserverMethod(AfterBeanDiscoveryImpl.java:366)
> ... 18 more
> 
> 
> We can see the AfterBeanDiscoveryImpl.java:180 being
> 
> /**
> * Bean and observer registration is delayed until after all {@link
> AfterBeanDiscovery} observers are notified.
> */
> private void finish() {
> try {
> GlobalEnablementBuilder globalEnablementBuilder =
> getBeanManager().getServices().get(GlobalEnablementBuilder.class);
> for (BeanRegistration registration : additionalBeans) {
> processBeanRegistration(registration, globalEnablementBuilder);
> }
> for (ObserverRegistration registration : additionalObservers) {
> processObserverRegistration(registration);
> }
> } catch (Exception e) {
> throw new DefinitionException(e); <<<<<< line 180
> }
> }
> 
> 
> I am also aware of, in http://docs.jboss.org/cdi/spec/2.0/cdi-spec.pdf
> 11.5.3.,
> "If any observer method of the AfterBeanDiscovery event throws an exception,
> the exception is treated as a definition error by the container"
> 
> So why is the test expecting DeploymentException
> 
> import javax.enterprise.inject.spi.De ploymentException;
> ... @ShouldThrowException(DeploymentException.class)
> 
> 
> It seems like the test is seeing
> @SpecAssertions({ @SpecAssertion(section = AFTER_BEAN_DISCOVERY, id = "ee")
> })
> If the container is unable to process the ObserverMethodConfigurator it
> automatically detects the problem and treats it as a deployment problem.
> as dominant but the weld implementation is seeing
> http://docs.jboss.org/cdi/spec/2.0/cdi-spec.pdf : 11.5.3
> If any observer method of the AfterBeanDiscovery event throws an exception,
> the exception is treated as a definition error by the container"
> as dominant.
> 
> ....and the two do not agree.
> 
> So the bottom line is the CDI CTS/TCK test
> org.jboss.cdi.tck.tests.extensions.lifecycle.broken.observerMethod.ObserverMethodWithoutNotifyMethodTest
> fails due to the inconsistency.
> 
> Have I missunderstood the spec contradiction or missed some part of Weld
> integration needed?
> 
> Gordon Hutchison
> 
> _______________________________________________
> weld-dev mailing list
> weld-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/weld-dev