How to use shouldDescribeWithExtraTypeInfoOfLong method of org.mockito.internal.matchers.MatchersToStringTest class

This should work

import static org.mockito.ArgumentMatchers.any;
import static org.mockito.Mockito.verify;

verify(bar).DoStuff(any(Foo[].class));

New feature added to Mockito makes this even easier,

ArgumentCaptor<Person> argument = ArgumentCaptor.forClass(Person.class);
verify(mock).doSomething(argument.capture());
assertEquals("John", argument.getValue().getName());

Take a look at Mockito documentation

In case when there are more than one parameters, and capturing of only single param is desired, use other ArgumentMatchers to wrap the rest of the arguments:

verify(mock).doSomething(eq(someValue), eq(someOtherValue), argument.capture());
assertEquals("John", argument.getValue().getName());

Also, a slightly more terse way to get around this is to use the do syntax instead of when.

doReturn(User.class).when(methodParameter).getParameterType();

Best method is to use WireMock. Add the following dependencies:

    <dependency>
        <groupId>com.github.tomakehurst</groupId>
        <artifactId>wiremock</artifactId>
        <version>2.4.1</version>
    </dependency>
    <dependency>
        <groupId>org.igniterealtime.smack</groupId>
        <artifactId>smack-core</artifactId>
        <version>4.0.6</version>
    </dependency>

Define and use the wiremock as shown below

@Rule
public WireMockRule wireMockRule = new WireMockRule(8089);

String response ="Hello world";
StubMapping responseValid = stubFor(get(urlEqualTo(url)).withHeader("Content-Type", equalTo("application/json"))
            .willReturn(aResponse().withStatus(200)
                    .withHeader("Content-Type", "application/json").withBody(response)));

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();

For these kinds of tests I like to use a little library called awaitility. You can do it yourself with a countdown latch, but as you have seen you have to hack up your test with a machete to make that work.

In this test you should call verify after you have awaited the latch.

Another problem in your code is the private Boolean response. Since you're changing it in another thread you should make that an AtomicBoolean or at least declare it volatile.

One way could be to avoid being too restrictive on your arguments in order to provide all the expected results with only one thenReturn call.

For example let's say that I want to mock this method:

public String foo(String firstArgument, Object obj) {
    return "Something";
}

You could then mock it by providing as many results as you want like below:

// Mock the call of foo of any String to provide 3 results
when(mock.foo(anyString(), anyObject())).thenReturn("val1", "val2", "val3");

Calls to foo with any parameters will provide respectively "val1", "val2", then any subsequent calls will provide "val3".

In case you do care about passed values but don't want to depend on call sequence you can use thenAnswer to provide an answer that matches with the second argument like you currently do but with 3 different thenReturn. Assuming that you have overridden the method equals(Object o).

when(mock.foo(anyString(), anyObject())).thenAnswer(
    invocation -> {
        Object argument = invocation.getArguments()[1];
        if (argument.equals(new ARequest(1, "A"))) {
            return new AResponse(1, "passed");
        } else if (argument.equals(new ARequest(2, "2A"))) {
            return new AResponse(2, "passed");
        } else if (argument.equals(new BRequest(1, "B"))) {
            return new BResponse(112, "passed");
        }
        throw new InvalidUseOfMatchersException(
            String.format("Argument %s does not match", argument)
        );
    }
);

Or simply, using the methods anyString and eq as argument marchers. Assuming that you have overridden the method equals(Object o).

when(service.foo(anyString(), eq(new ARequest(1, "A"))))
    .thenReturn(new AResponse(1, "passed"));
when(service.foo(anyString(), eq(new ARequest(2, "2A"))))
    .thenReturn(new AResponse(2, "passed"));
when(service.foo(anyString(), eq(new BRequest(1, "B"))))
    .thenReturn(new BResponse(112, "passed"));

For some reason, your test suite, tries to load the MockitoJunitRunner from the org.mockito.junit contained in the Mockito versions >= 2. O. In that version, the line:

at org.mockito.junit.MockitoJUnitRunner.<init>(MockitoJUnitRunner.java:152)

is doing this:

public MockitoJUnitRunner(Class<?> klass) throws InvocationTargetException {
        //by default, StrictRunner is used. We can change that potentially based on feedback from users
        this(new StrictRunner(new RunnerFactory().createStrict(klass), klass));
    }

and the RunnerFactory that is loaded here is from version 1.x as createStrict has been introduced in Mockito 2.x.

So go through the pom dependency tree and to find what artifact implicitly add the Mockito 2.x dependency to your project and exclude it.

Alternatively.. as a workaround, instead of the @RunWith(MockitoJUnitRunner.class) you can use:

@Before
public void init() {
    MockitoAnnotations.initMocks(this);
}

You can also check out this Mockito cheat sheet to keep all the standards at hand.

For functions returning void, use doAnswer()

doAnswer(...).when(mockedObject).handle(any(Callback[].class));

And an Answer that performs the interception must go in as the parameter to doAnswer, e.g. as an anonymous class:

new Answer() {
  public Object answer(InvocationOnMock invocation) {
      Object[] args = invocation.getArguments();
      Mock mock = invocation.getMock();
      return null;
  }}

In this case args will be the array Callback[]!