How to use BulkEmailService method of org.mockitousage.matchers.CustomMatcherDoesYieldCCETest class

Okey, found the answer, you need to call to

@PrepareForTest(Foo.class)

instead of

@PrepareForTest(Bar.class)

It seems Mockito gets confused when you call your deep stub in then when method. I was able to work around it by calling mock.getBar() separately:

    Foo mock = mock(Foo.class, RETURNS_DEEP_STUBS);
    Bar bar = mock.getBar();
    doReturn("deep").when(bar).getName();

@Spy doc says:

A field annotated with @Spy can be initialized explicitly at declaration point. Alternatively, if you don't provide the instance Mockito will try to find zero argument constructor (even private) and create an instance for you.

@SpyBean doc says:

Annotation that can be used to apply Mockito spies to a Spring ApplicationContext.

All beans in the context of the same type will be wrapped with the spy. If no existing bean is defined a new one will be added.

So the main difference is @SpyBean is a Spring Boot specific annotation but @Spy is part of Mockito itself. @SpyBean and @Spy basically do the same, but @SpyBean can resolve the Spring specific dependencies, e.g. @Autowired, @Spy can only create object with empty constructor.

Use an Answer to check the value of the argument when the method is called. You can either throw an AssertionError within the Answer if the value is wrong, or you can store the value, and do your assertion at the end.

On a first glance, the answer is simply: well, there are several mocking frameworks out there, and there are different ways to use them.

The first example tells JUnit to use the "unit test runner" that the Mockito mocking framework provides. The second example uses the unit test runner from the PowerMock framework.

In order for things to make sense, you would also have different import statements, as both frameworks have different implementations for the @Mock annotation for example.

( the main point of using these framework-specific test runners is that they take care of initializing all the fields with special framework-specific annotations ).

So: the difference here is simply that: the first example is written to use the Mockito framework, the second one uses PowerMock.

Now, which one of those to use?

Answer: Mockito.

Why? Somehow an ugly truth is: the PowerMock-one basically is a cry for help. It says "the class under test is badly designed, please fix it". Meaning: as a developer, you can write "easy to test" code, or "hard to test" code. Many people do the second: they write code that is hard to test. And then, PowerMock(ito) provides means to still test that code.

PowerMock(ito) gives you the ability to mock (thus control) calls to static methods, and to new(). To enable that, PowerMock(ito) manipulates the byte code of your code under test. That is perfectly fine for small code bases, but when you face millions of lines of production code, and thousands of unit tests, things are totally different.

I have seen many PowerMock tests fail for no apparent reason, to find out hours later ... that some "static" thing somewhere else was changed, and that somehow affect a different PowerMock static/new driven test case.

At some point, our team made a conscious decision: when you write new code, and you can only test that with PowerMock ... that isn't acceptable. Since then, we only created Mockito test cases, and not once since then we saw similar bizarre problems that bugged us with PowerMock.

The only acceptable reason to use PowerMock is when you want to test existing (maybe 3rd party) code that you do not want to modify. But of course, what is the point of testing such code? When you can't modify that code, why should tests fail all of a sudden?

Mockito 1.10.5 has introduced this feature.

For the code sample in the question, here is one way to capture the varargs:

    ArgumentCaptor<String> varArgs = ArgumentCaptor.forClass(String.class);
    A mock = Mockito.mock(A.class);
    mock.setNames("Jeff", "Mike", "John");
    Mockito.verify(mock).setNames(varArgs.capture());

    //Results may be validated thus:
    List<String> expected = Arrays.asList("Jeff", "Mike", "John");
    assertEquals(expected, varArgs.getAllValues());

Please see the ArgumentCaptor javadoc for details.

Handmade

As InitialContext doc says, you can provide your own factory for InitialContext objects, using java.naming.factory.initial system property. When the code runs inside application server, the system property is set by the server. In our tests, we provide our own implementation of JNDI.

Here's my Mockito only solution: I defined a custom InitialContextFactory class, that returns a mock of InitialContext. You customize the mock as you wish, probably to return more mocks on lookup calls.

public class PlainTest {
  @Mock InitialContextFactory ctx;
  @InjectMocks Klasa1 klasa1;

  public static class MyContextFactory implements InitialContextFactory
  {
    @Override
    public Context getInitialContext(Hashtable<?, ?> environment) throws NamingException {
      ConnectionFactory mockConnFact = mock(ConnectionFactory.class);
      InitialContext mockCtx = mock(InitialContext.class);
      when(mockCtx.lookup("jms1")).thenReturn(mockConnFact);
      return mockCtx;
    }
  }

  @Before
  public void setupClass() throws IOException
  {
    MockitoAnnotations.initMocks(this);
    System.setProperty("java.naming.factory.initial",
      this.getClass().getCanonicalName() + "$MyContextFactory");
  }

Spring (added by edit)

If you don't mind leveraging Spring Framework for testing purposes, here's their simple solution: SimpleNamingContextBuilder:

SimpleNamingContextBuilder builder = new SimpleNamingContextBuilder();
DataSource ds = new DriverManagerDataSource(...);
builder.bind("java:comp/env/jdbc/myds", ds);
builder.activate();

It's ok to put it in @Before or @BeforeClass. After activate(), jndi data will be pulled from spring dummy.

The junit5-migration-gradle project demonstrates how to execute tests based on JUnit 5 using Gradle. In addition, it showcases that existing JUnit 4 based tests can be executed in the same test suite as JUnit Jupiter based tests or any other tests supported on the JUnit Platform.

Basically it boils down to having both engines, Jupiter and Vintage, on the runtime class-path:

dependencies {
    testCompile("org.junit.jupiter:junit-jupiter-api:5.2.0")
    testRuntime("org.junit.jupiter:junit-jupiter-engine:5.2.0")
}

dependencies {
    testCompile("junit:junit:4.12")
    testRuntime("org.junit.vintage:junit-vintage-engine:5.2.0")
}

If the contract of class A includes the fact that it calls method B of an object of type C, then you should test this by making a mock of type C, and verifying that method B has been called.

This implies that the contract of class A has sufficient detail that it talks about type C (which might be an interface or a class). So yes, we're talking about a level of specification that goes beyond just "system requirements", and goes some way to describing implementation.

This is normal for unit tests. When you are unit testing, you want to ensure that each unit is doing the "right thing", and that will usually include its interactions with other units. "Units" here might mean classes, or larger subsets of your application.

Update:

I feel that this doesn't apply just to verification, but to stubbing as well. As soon as you stub a method of a collaborator class, your unit test has become, in some sense, dependent on implementation. It's kind of in the nature of unit tests to be so. Since Mockito is as much about stubbing as it is about verification, the fact that you're using Mockito at all implies that you're going to run across this kind of dependency.

In my experience, if I change the implementation of a class, I often have to change the implementation of its unit tests to match. Typically, though, I won't have to change the inventory of what unit tests there are for the class; unless of course, the reason for the change was the existence of a condition that I failed to test earlier.

So this is what unit tests are about. A test that doesn't suffer from this kind of dependency on the way collaborator classes are used is really a sub-system test or an integration test. Of course, these are frequently written with JUnit too, and frequently involve the use of mocking. In my opinion, "JUnit" is a terrible name, for a product that lets us produce all different types of test.