Running concurrent test fails with java.util.EmptyStackException

[krop-freenetdigital]

Using the JBehave feature of running tests concurrently fails under certain circumstances while writing the Thucydides report with the following exception:

[WARNING] Failed to run story stories/booking/Unlock.story

java.lang.RuntimeException: java.lang.reflect.InvocationTargetException
at org.jbehave.core.reporters.ConcurrentStoryReporter$DelayedMethod.invoke(ConcurrentStoryReporter.java:356)
at org.jbehave.core.reporters.ConcurrentStoryReporter.invokeDelayed(ConcurrentStoryReporter.java:335)
at org.jbehave.core.embedder.StoryRunner.run(StoryRunner.java:230)
at org.jbehave.core.embedder.StoryRunner.run(StoryRunner.java:181)
at org.jbehave.core.embedder.StoryManager$EnqueuedStory.call(StoryManager.java:262)
at org.jbehave.core.embedder.StoryManager$EnqueuedStory.call(StoryManager.java:229)
at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:334)
at java.util.concurrent.FutureTask.run(FutureTask.java:166)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1110)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:603)
at java.lang.Thread.run(Thread.java:722)
Caused by: java.lang.reflect.InvocationTargetException
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:601)
at org.jbehave.core.reporters.ConcurrentStoryReporter$DelayedMethod.invoke(ConcurrentStoryReporter.java:352)
... 10 more
Caused by: java.util.EmptyStackException
at java.util.Stack.peek(Stack.java:102)
at java.util.Stack.pop(Stack.java:84)
at net.thucydides.jbehave.ThucydidesReporter.afterScenario(ThucydidesReporter.java:532)
at org.jbehave.core.reporters.DelegatingStoryReporter.afterScenario(DelegatingStoryReporter.java:43)
... 15 more

It is difficult to reproduce because it does not happen on any computer. In our case only to occurs when at least on scenario of a story is failing and multiple stories are used. Even than, it does not happen on every host - because a race condition seems to be involved, but it can be validated by setting some break points. It seems that net.thucydides.jbehave.ThucydidesReporter.activeScenarios is responsible for the failure, which is accessed concurrently by multiple threads.

We have tried out a fix where activeScenarios is accessed via a thread local variable which worked fine on our machines.

Best regards

Tronje