NullPointerException in SpringTransactionHooks when upgrading to 1.1.5

[mmalmeida]

Running the tests after upgrading from 1.1.3 to 1.1.5 yields the following exception:

java.lang.NullPointerException
    at cucumber.api.spring.SpringTransactionHooks.obtainPlatformTransactionManager(SpringTransactionHooks.java:65)
    at cucumber.api.spring.SpringTransactionHooks.rollBackTransaction(SpringTransactionHooks.java:60)
    at sun.reflect.GeneratedMethodAccessor17.invoke(Unknown Source)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
    at java.lang.reflect.Method.invoke(Method.java:597)
    at cucumber.runtime.Utils$1.call(Utils.java:35)
    at cucumber.runtime.Timeout.timeout(Timeout.java:12)
    at cucumber.runtime.Utils.invoke(Utils.java:31)
    at cucumber.runtime.java.JavaHookDefinition.execute(JavaHookDefinition.java:59)
    at cucumber.runtime.Runtime.runHookIfTagsMatch(Runtime.java:213)
    at cucumber.runtime.Runtime.runHooks(Runtime.java:201)
    at cucumber.runtime.Runtime.runAfterHooks(Runtime.java:195)
    at cucumber.runtime.model.CucumberScenario.run(CucumberScenario.java:42)
    at cucumber.runtime.junit.ExecutionUnitRunner.run(ExecutionUnitRunner.java:83)
    at cucumber.runtime.junit.FeatureRunner.runChild(FeatureRunner.java:63)
    at cucumber.runtime.junit.FeatureRunner.runChild(FeatureRunner.java:18)
    at org.junit.runners.ParentRunner$3.run(ParentRunner.java:238)
    at org.junit.runners.ParentRunner$1.schedule(ParentRunner.java:63)
    at org.junit.runners.ParentRunner.runChildren(ParentRunner.java:236)
    at org.junit.runners.ParentRunner.access$000(ParentRunner.java:53)
    at org.junit.runners.ParentRunner$2.evaluate(ParentRunner.java:229)
    at org.junit.runners.ParentRunner.run(ParentRunner.java:309)
    at cucumber.runtime.junit.FeatureRunner.run(FeatureRunner.java:70)
    at cucumber.api.junit.Cucumber.runChild(Cucumber.java:82)
    at cucumber.api.junit.Cucumber.runChild(Cucumber.java:41)
    at org.junit.runners.ParentRunner$3.run(ParentRunner.java:238)
    at org.junit.runners.ParentRunner$1.schedule(ParentRunner.java:63)
    at org.junit.runners.ParentRunner.runChildren(ParentRunner.java:236)
    at org.junit.runners.ParentRunner.access$000(ParentRunner.java:53)
    at org.junit.runners.ParentRunner$2.evaluate(ParentRunner.java:229)
    at org.junit.runners.ParentRunner.run(ParentRunner.java:309)
    at cucumber.api.junit.Cucumber.run(Cucumber.java:87)
    at org.eclipse.jdt.internal.junit4.runner.JUnit4TestReference.run(JUnit4TestReference.java:50)
    at org.eclipse.jdt.internal.junit.runner.TestExecution.run(TestExecution.java:38)
    at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:467)
    at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.runTests(RemoteTestRunner.java:683)
    at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.run(RemoteTestRunner.java:390)
    at org.eclipse.jdt.internal.junit.runner.RemoteTestRunner.main(RemoteTestRunner.java:197)

Originally posted in https://groups.google.com/forum/#!topic/cukes/SjFK1QW2qE0

[mmalmeida]

I believe this might be due to the fact that the class does not have @ContextConfiguration("classpath:cucumber.xml") . If you create a completely equal class but apply the ContextConfiguration to it, the error will disappear.

Two related comments:

1. Writing @ContextConfiguration("classpath:cucumber.xml") on every stepdef and hook seems very cumbersome. In addition, if you need to define the cucumber.xml file on the hook the hook will be limited to usages with that specific file name.

2. This might not have been caught because I believe the example in examples/spring-txn/src/test/java/cucumber/examples/spring/txn/RunCukesTest.java is wrong - the hook should be "cucumber.api.spring" and not "cucumber.runtime.java.spring.hooks"

[pasviegas]

hey @ffbit do you know anything about the spring-txn example being broken?

[mmalmeida]

@pasviegas, @ffbit - The test on master is passing - but this is due to the fact that the Hook is not being called (the path in glue is wrong, it was changed in 1.1.1). See my branch above, where the hook path was corrected and the test consequently fails.

[anaMZ]

Hello,
I have the same error. Have you fixed it or do you know a workaround?
Thank you!

[ffbit]

I'm on vocation without my laptop (I'm back on Monday).
I ran all the examples when the #632 and #633 PR were made. Only the wicket example was red.

I MUST write the Spring migration guide, I promised that this summer, but haven't done yet.

[mmalmeida]

@ffbit - did you run the examples with the correct spring hook path, though (as in the commit I posted above)? As this definitely makes the test fail!

[ffbit]

I can't answer this question until I get to my laptop next Monday.

[brasmusson]

Hi, @ffbit and others, I wrapped the glue path change of the spring-txn example in a pull request (#644), so now the error in the example is clearly shown on the two Travis builds that run the examples (openidk6 and openjdk7).

[ffbit]

@brasmusson thanks for the clarification

[ffbit]

Hi, I've been trying to fix the issue and here are my findings.

I got rid of the NPE, by turning the c.a.s.SpringTransactionHooks into a Spring bean.

Then I faced another problem

org.springframework.beans.factory.NoSuchBeanDefinitionException:
  No qualifying bean of type [org.springframework.transaction.PlatformTransactionManager] is defined
...

@pasviegas, could you please correct me if I went wrong.
I found that there are two separate application context types in:

• c.r.j.s.SpringFactory
• every step definition class marked with a @*Configuration annotation.

The first one is neither a parent nor a child for the second one and they live their own lives.
So, in the first case SpringTransactionHooks's application context can't resolve the needed PlatformTransactionManager dependency, and in the second case the needed dependency is present but unreachable.

As a summary I can send a pull request solving the NPE and we can start solving the NoSuchBeanDefinitionException.

WDYT?

[pasviegas]

No, this shouldn't be the case.
Every execution should be using the same context and cleaning it when it finishes.
I will take a look as soon as I get home. Can you open a branch an send the PR?