Let's say I have a test class called testFixtureA with several methods testA, testB, testC, etc, each with #Test annotation.
Let's now say I subclass testFixtureA into class called testFixtureAB and I don't overwrite anything. testFixtureAB is empty as for now.
When I run tests from testFixtureAB, methods testA, testB and testC are executed by test runner because test runner doesn't distinguish between test methods from class and baseclass.
How can I force test runner to leave out tests from baseclass?
and I don't overwrite anything.
testFixtureAB is empty as for now
There's your answer. If you want to not run testB from the main class, overrride it:
public class testFixtureAB extends testFixtureA {
#Override
public void testB() {}
}
Restructure your test classes.
If you don't want to use the tests from the baseclass, then don't extend it
If you need other functionality from the base class, split that class in two - the tests, and the other functionality
ignoring the whole base class:
#Ignore
class BaseClass {
// ...
}
check out this example
It's quite easy to achieve implementing some few classes:
Create your own TestRunner
Create an annotation like #IgnoreInheritedTests
Create a class that extends org.junit.runner.manipulation.Filter
On the filter class:
public class InheritedTestsFilter extends Filter {
#Override
public boolean shouldRun(Description description) {
Class<?> clazz = description.getTestClass();
String methodName = description.getMethodName();
if (clazz.isAnnotationPresent(IgnoreInheritedTests.class)) {
try {
return clazz.getDeclaredMethod(methodName) != null;
} catch (Exception e) {
return false;
}
}
return true;
}
#Override
public String describe() {
// TODO Auto-generated method stub
return null;
}
}
on your custom runner:
/**
* #param klass
* #throws InitializationError
* #since
*/
public CustomBaseRunner(Class<?> klass) throws InitializationError {
super(klass);
try {
this.filter(new InheritedTestsFilter());
} catch (NoTestsRemainException e) {
throw new IllegalStateException("class should contain at least one runnable test", e);
}
}
In Junit 5, you can make base class as abstract and extends it with a concrete class.
When you run the abstract in your IDE, your subclass will get executed instead.
In the latest JUnit you can use the #Rule annotation on the subclass to inspect the test name and intercept the test run to ignore the test dynamically. But I would suggest that #Bozho's idea is the better one - the fact that you need to do this indicates a bigger problem that probably shows inheritance is not the right solution here.
I know, it's not the answer...
Consider the reason why you extend concrete test classes. You do duplicate test methods that way.
If you share code between tests then consider writing base test classes with helper and fixture setup methods or test helper class.
If for running tests then try organizing tests with suites and categories.
What if you want to execute the same test for different configurations of the same test suite?
For example, let's say you have Class A with test1, test2 and test3 methods that hit an embedded database then you want to create separated "setUp" and "tearDown" for every embedded vendor (H2, HyperSQL, etc) but runs the same testing for each one.
I would like to extend a class that contain those test methods and configure it in a subclass. My problem is that the super class SHOULD NOT be considered as eligible for the test runner. The problem arises when the test runner executes the super class and given that don't found the corresponding setup and teardown methods, it crashs :(
In the base test class' #Test methods:
assumeTrue(getClass().equals(BaseClassTest.class));
It will ignore those in the subclass tests but not completely leave them out.
If for any reason you need two JUnit classes for same functionality, the best approach for me is:
put the common code in a parent class TestFixture with only constants and mocked services.
create two subclasses: TestFixtureA and TestFixtureB
This way you will not have duplicated code, nor double runs.
I ran into a similar issue where I had a generic abstract base class TestBase and 2 subclasses TestA, TestB that extended the base class. I wanted all tests from TestBase to run for TestA but to skip some for TestB. You can use the JUnit annotations Nested and Disabled to group tests in the base class into categories and then enable a full disable of that entire category of tests.
public class TestB extends TestBase {
// disable entire suite of inherited tests
// from the nested class declared on the parent
#Nested
#Disabled
public class ComponentCTests {}
}
public abstract class TestBase {
#Nested
public class ComponentCTests {
#Test
public void testSomething() {
// test something
}
}
}
Related
I'm using PowerMock to mock a java.net.Inet4Address (amongst other things) to return a particular IP address (getHostAddress()), as well as whether or not it's loopback (isLoopbackAddress()). I find that if I perform the actual stubbing (PowerMock.doReturn(...).when(mock).methodToStub()) anywhere other than from within the test class or an immediately inner class I get an UnfinishedStubbingException.
The problem is most obvious if I try to perform two stubs. The first one passes without an exception, but the second one throws the exception because it thinks the first one was not finished. If I only perform the one stub then I see different errors depending on what I do after that, so it's definitely the first stub that's causing the problem.
Below is some code that demonstrates the problem.
TestClass.java
/* package, imports... */
#RunWith(PowerMockRunner.class)
#PrepareForTest({Inet4Address.class})
public class TestClass {
#Test
public void test() {
Inet4Address mocked = PowerMockito.mock(Inet4Address.class);
// Option 1: Do it from within this class - WORKS
doStubbing(mocked);
// Option 2: Do it from an inner class - WORKS
Inner.doStubbing(mocked);
// Option 3: Do it from an inner class of the inner class - FAILS
Inner.Deeper.doStubbing(mocked);
// Option 4: Do it from an entirely different class - FAILS
OtherClass.doStubbing(mocked);
}
private void doStubbing(Inet4Address mocked) {
PowerMockito.doReturn(true).when(mocked).isLoopbackAddress();
PowerMockito.doReturn("127.0.0.1").when(mocked).getHostAddress();
}
public static class Inner {
static void doStubbing(Inet4Address mocked) {
PowerMockito.doReturn(true).when(mocked).isLoopbackAddress();
PowerMockito.doReturn("127.0.0.1").when(mocked).getHostAddress();
}
public static class Deeper {
static void doStubbing(Inet4Address mocked) {
PowerMockito.doReturn(true).when(mocked).isLoopbackAddress();
PowerMockito.doReturn("127.0.0.1").when(mocked).getHostAddress();
}
}
}
}
OtherClass.java
/* package, imports... */
public class OtherClass {
public static void doStubbing(Inet4Address mocked) {
PowerMockito.doReturn(true).when(mocked).isLoopbackAddress();
PowerMockito.doReturn("127.0.0.1").when(mocked).getHostAddress();
}
}
I've put the creation of the mock at the start, common to all scenarios. It makes no difference if the mock is created from within the same class where the stubbing is being done. I've also made the methods static for ease of reading; the behaviour is the same if the classes are instantiated first.
I know there are workarounds so I can get my test working (perform the mocking right there in the test class, mock the InetAddress interface instead of the IPv4 implementation, etc) but I'd like to know why PowerMock is behaving in this way. I could almost understand it if it only worked from within the test class, but why does it work in an inner class as well?
My question is about testing a class that implements many interfaces. For example, I have this class:
public class ServiceControllerImpl extends ServiceController implements IDataChanged, IEventChanged {
}
Now there are two ways for testing. The first is testing directly on the concrete class. That means the object type is the concrete class rather than the interface.
public class ServiceControllerImplTest {
ServiceControllerImpl instance;
#Before
public void setUp() {
instance = new ServiceControllerImpl();
// you can bring this instance anywhere
}
}
The second way is testing on the interface only. We must typecast this object to all interfaces it implements.
public class ServiceControllerImplTest {
ServiceController instance; // use interface here
IDataChanged dataChangeListener;
#Before
public void setUp() {
instance = new ServiceControllerImpl();
dataChangeListener = (IDataChanged) instance;
// instance and dataChangeListener "look like" two different object.
}
}
I prefer the second solution because maybe in future we can change the interface it implements to other objects, so using the concrete class might lead to failing tests in the future. I don't know the best practice for this problem.
Thanks :)
I prefer second solution because in reality, maybe in future we can change the interface it implements to other objects, so force using concreted class maybe leads to fail test in the future.
I guess it will lead to failed tests anyway, because you usually test that assertions are true or false. The question is: Do that tests apply to any IDataChanged or do these assertions only apply to the ServiceControllerImpl?
If the assertions only apply to the ServiceControllerImpl it doesn't matter if you use an IDataChanged instead of an ServiceControllerImpl, because you must edit the test when you use another IDataChanged object - different assertions. The test will fail if you use another object.
The way you setup unit tests Itself gives you an answer. A unit test usually tests one class in isolation. This means that you mock the environment. But mocking the environment means that you know the dependencies of the class you test and this are implementation details. So your test is written on an implemtation basis rather than only the interface.
It's possible to write tests that only test an abstract api - like an interface. But this usually means that your tests are abstract too. E.g.
public abstract class SetTest {
#Test
public void addAlreadyExistentObject(){
Set<String> setUnderTest = createSetUnderTest();
Assert.assertTrue(setUnderTest.isEmpty());
boolean setChanged = setUnderTest.add("Hello");
Assert.assertTrue(setChanged);
setChanged = setUnderTest.add("Hello");
Assert.assertFalse(setChanged);
Assert.assertEquals(setUnderTest.size(), 1);
}
protected abstract Set<String> createSetUnderTest();
}
You can then extend these abstract tests to test the api for concrete classes. E.g.
public class HashSetTest extends SetTest {
#Override
protected Set<String> createSetUnderTest() {
return new HashSet<String>();
}
}
In this case you can replace the implementation and the test must remain green.
But here is another example of an abstract api when replacing the object under test does not really make sense.
What about writing a test for all Runnables?
public class RunnableTest {
#Test
public void run(){
Runnable runnable = ...;
// What to test here?
// run is invoked without throwing any runtime exceptions?
runnable.run();
}
}
As you can see it does not make sense in some cases to write tests in a way so that you can easily replace the object under test.
If an api like the Set api defines a concrete state handling you can write abstract tests that test this.
JayC667 already correctly answered that it's best to refer to a class through its supertype(s) in tests of methods defined by those types. But I'd change the way you did that a bit to avoid casting:
public class ServiceControllerImplTest {
ServiceController controller;
IDataChanged dataChangeListener;
#Before
public void setUp() {
instance = new ServiceControllerImpl();
controller = instance;
dataChangeListener = instance;
}
}
I have a class, which I use as a basis for my unit tests. In this class I initialize the whole environment for my tests, setting up database mappings, enter a number of database records across multiple tables, etc. That class has a method with a #BeforeClass annotation which does the initialization. Next thing, I extend that class with specific classes in which I have #Test methods.
My question is, since the before class is exactly the same for all these test classes, how can I ensure that they are run only once for all the tests.
One simple solution is that I could keep all the tests in one class. However, the number of tests is huge, also they are categorised based on functional heads. So they are located in different classes. However since they need the exact same setup, they inherit the #BeforeClass. As a result the whole setup is done at least once per test class, taking much more time in total than I would prefer.
I could, though, put them all in various subpackages under one package, hence if there is a way, how I can run set up once for all the tests within that package, it would be great.
With JUnit4 test suite you can do something like this :
#RunWith(Suite.class)
#Suite.SuiteClasses({ Test1IT.class, Test2IT.class })
public class IntegrationTestSuite
{
#BeforeClass
public static void setUp()
{
System.out.println("Runs before all tests in the annotation above.");
}
#AfterClass
public static void tearDown()
{
System.out.println("Runs after all tests in the annotation above.");
}
}
Then you run this class as you would run a normal test class and it will run all of your tests.
JUnit doesn't support this, you will have to use the standard Java work-arounds for singletons: Move the common setup code into a static code block and then call an empty method in this class:
static {
...init code here...
}
public static void init() {} // Empty method to trigger the execution of the block above
Make sure that all tests call init(), for example my putting it into a #BeforeClass method. Or put the static code block into a shared base class.
Alternatively, use a global variable:
private static boolean initialize = true;
public static void init() {
if(!initialize) return;
initialize = false;
...init code here...
}
Create one base class for all tests:
public class BaseTest {
static{
/*** init code here ***/
}
}
and every test should inherit from it:
public class SomeTest extends BaseTest {
}
You can make one BaseTest class with a #BeforeClass method, then have all the other tests inherit from it. This way, when each test object is constructed, #BeforeClass gets executed.
Also avoid executing it just once for all the test suite, since all the test cases should be independent. #BeforeClass should execute only once each test case, not test suite.
If you can tolerate adding spring-test to your project, or you are using it already, then a good approach is to use the technique described here: How to load DBUnit test data once per case with Spring Test
Not sure if anyone still is using JUnit and trying to fix it without using Spring Runner (aka no spring integration). TestNG has this feature. But here is a JUnit based solution.
Create a RunOnce per thread operation like so. This maintains a list of classes for which the operation has run.
public class RunOnceOperation {
private static final ThreadLocal t = new ThreadLocal();
public void run(Function f) {
if (t.get() == null) {
t.set(Arrays.asList(getClass()));
f.apply(0);
} else {
if (!((List) t.get()).contains(getClass())) {
((List) t.get()).add(getClass());
f.apply(0);
}
}
}
}
Back in your unit test
#Before
public beforeTest() {
operation.run(new Function<Integer, Void>() {
#Override
public Void apply(Integer t) {
checkBeanProperties();
return null;
}
});
}
private void checkBeanProperties() {
//I only want to check this once per class.
//Also my bean check needs instance of the class and can't be static.
}
My function interface is like this:
interface Function<I,O> {
O apply(I i);
}
When you use this way, you can perform operations once per class using ThreadLocal.
We have DAO tests that should run against both the real DAO/database, and against a mock dao to verify that the mock dao behaves the same as the real dao. To this end, we have a structure like this:
public abstract class DAOTestBase
{
public void testSimple()
{
// dummy assertion
assertTrue(true, "Hello");
}
}
#Test(groups = "fast")
public class TestMockDAO extends DAOTestBase
{
// setUp/tearDown and helper methods for mock
}
#Test(groups = "slow")
public class TestDAO extends DAOTestBase
{
// setUp/tearDown and helper methods for real DB
}
Unfortunately this doesn't work - TestNG doesn't think that the testSimple method is a test and hence won't run it. So instead I tried to annotate the testSimple method (or the DAOTestBase class):
A #Test annotation without any groups will lead to the same effect - the test won't run for either fast nor slow groups.
A #Test annotation with groups fast and slow will lead to the opposite effect - both TestMockDAO and TestDAO will be run regardless of whether only fast or only slow tests should be run.
A #Test annotation with a different group, say common, plus added dependsOnGroups="common" annotations in both TestMockDAO and TestDAO will also not work unless common is included in the groups to run which leads again to case 2 above (both TestMockDAO and TestDAO are run).
In the end, what I'm looking for is a way to be able to define the group for the inherited tests in the sub class, but it seems as if the #Test annotation is only applied to test methods in that very same class, not also to inherited methods that don't have a #Test annotation. Is there any other way to achieve this (without overriding all methods in the sub classes) ?
I am currently working through a similar situation.
A way to make test cases run is to use something like:
#Test
public void someTest() {
TestNG testng = new TestNG();
testng.setTestClasses(new Class[] { SomeTests.class });
testng.run();
}
Reference: http://testng.org/doc/documentation-main.html#running-testng-programmatically
Unforunately I am currently unable to get it to report the test cases within SomeTests.
Have you tried simply adding a #Test annotation on top of DAOTestBase? Each subclass will override it with its own group and this should make the method in the base a test method.
I am using TestNG 6.14.3 version and I found a solution using priority annotation.
Example:
I have a base test class:
public class TestBase {
#Test(priority = 0)
public void testA() {
assertTrue(true, "testA");
}
}
And another extended test class:
public class Test2 extends TestBase {
#Test(priority = 1)
public void testB() {
assertTrue(true, "testB");
}
}
When I run Test2 test class, I obtain the following esult:
testA: true
testB: true
I solved it this way:
The methods in the base class are in "base" group, but need to check if the test have been initialized.
public abstract DaoTestBase {
private boolean initialized = false;
#Test(groups = "base")
public void testSimple() {
if (!initialized) { return; }
// dummy assertion
assertTrue(true, "Hello");
}
}
The test is initialized in the child, in the BeforeClass annotated method.
#BeforeClass
protected void initialize() {
super.initialized = true;
}
If you annotate the parent class instead of the methods, you must pass inheritGroups=false and the group, since it inherits also the group of the base class and it will not work.
Now, you must run TestNG to check groups base,fast or base,slow. Both tests will be executed, but the one not initialized will do nothing.
It is ugly, and I would not recommend it (it looks better to redefine the methods in child and call the according super method), but in my case I need priority in my test methods, and I want to avoid that repetition in each child class.
I have an abstract base class, which I use as a base for my unit tests (TestNG 5.10). In this class, I initialize the whole environment for my tests, setting up database mappings, etc. This abstract class has a method with a #BeforeClass annotation which does the initialization.
Next, I extend that class with specific classes in which I have #Test methods and also #BeforeClass methods. These methods do class-specific initialization of the environment (e.g. put some records into the database).
How I can enforce a specific order of the #BeforeClass annotated methods? I need the ones from the abstract base class to be executed before the ones of the extending class.
Example:
abstract class A {
#BeforeClass
doInitialization() {...}
}
class B extends A {
#BeforeClass
doSpecificInitialization() {...}
#Test
doTests() {...}
}
Expected order:
A.doInitialization
B.doSpecificInitialization
B.doTests
Actual order:
B.doSpecificInitialization // <- crashes, as the base init is missing
(A.doInitialization // <---not executed
B.doTests) // <-/
edit: Answer below is for JUnit, but I will leave it here anyway, because it could be helpful.
According to the JUnit api: "The #BeforeClass methods of superclasses will be run before those the current class."
I tested this, and it seems to work for me.
However, as #Odys mentions below, for JUnit you need to have the two methods named differently though as doing otherwise will result in only the subclass method being run because the parent will be shadowed.
Don't put the #BeforeClass on the abstract class. Call it from each subclass.
abstract class A {
void doInitialization() {}
}
class B extends A {
#BeforeClass
void doSpecificInitialization() {
super.doInitialization();
}
#Test
void doTests() {}
}
Seems like TestNG has #BeforeClass(dependsOnMethods={"doInitialization"}) - give it a try.
I added public to the abstract class and TestNG (6.0.1) executed the doInitialization() before doTests. TestNG does not execute doInitialization() if I remove public from class A.
public abstract class A {
#BeforeClass
doInitialization() {...}
}
class B extends A {
#Test
doTests() {...}
}
I just tried your example with 5.11 and I get the #BeforeClass of the base class invoked first.
Can you post your testng.xml file? Maybe you are specifying both A and B there, while only B is necessary.
Feel free to follow up on the testng-users mailing-list and we can take a closer look at your problem.
--
Cedric
I've just gone through this and found one more way to achieve this. Just use alwaysRun on #BeforeClass or #BeforeMethod in the abstract class, works as you would expect.
public class AbstractTestClass {
#BeforeClass(alwaysRun = true)
public void generalBeforeClass() {
// do stuff
specificBeforeClass();
}
}
For JUnit:
As #fortega has mentioned:
According to the JUnit api: "The #BeforeClass methods of superclasses will be run before those the current class."
But be careful not to name both methods with the same name. Since in this case the parent method will be hidden by child parent. Source.
When I run from: JUnitCore.runClasses(TestClass.class);
It will execute the parent properly, before the child (You do not need super.SetUpBeforeClass();)
If you run it from Eclipse:
For some reason it fails to run the base class.
The work around:
Call the base class explicitely: (BaseTest.setUpBeforeClass();)
You may want to have a flag in the base class in case you run it from an application, to determine if it is already setup or not. So it only runs once if you run it via both possible methods (such as from eclipse for personal testing, and through ANT for a build release).
This appears to be a bug with Eclipse, or at least unexpected results..
How about having your #BeforeClass method call an empty specificBeforeClass() method that may or may not be overwritten by sub classes like so:
public class AbstractTestClass {
#BeforeClass
public void generalBeforeClass() {
// do stuff
specificBeforeClass();
}
protected void specificBeforeClass() {}
}
public class SpecificTest {
#Override
protected void specificBeforeClass() {
// Do specific stuff
}
// Tests
}
dependsOnMethod can be used.
e.g. in case of Spring (AbstractTestNGSpringContextTests)
#BeforeClass(alwaysRun = true, dependsOnMethods = "springTestContextPrepareTestInstance")
Check your import statement.
It should be
import org.testng.annotations.BeforeClass;
not
import org.junit.BeforeClass;
This works for me --
abstract class A {
#BeforeClass
doInitialization() {...}
}
class B extends A {
#Override
#BeforeClass
doInitialization() {
//do class specific init
}
#Test
doTests() {...}
}
Why don't you try to create an abstract method doSpecialInit() in your super class, called from your BeforeClass annotated method in superclass.
So developpers inheriting your class is forced to implement this method.
There is another easy solution here.
My particular situation is that I need to inject mock services from "BeforeClass" in the subclass before "BeforeClass" in the superclass is executed.
To do this - simply use a #ClassRule in the subclass.
For example:
#ClassRule
public static ExternalResource mocksInjector = new ExternalResource() {
#Override
protected void before() {
// inject my mock services here
// Note: this is executed before the parent class #BeforeClass
}
};
I hope this helps. This can effectively execute static setup in "reverse" order.
I've faced a similar issue today, the only difference was a Base class was not abstract
Here's my case
public class A {
#BeforeClass
private void doInitialization() {...}
}
public class B extends A {
#BeforeClass
private void doSpecificInitialization() {...}
#Test
public void doTests() {...}
}
It occurred that a #BeforeClass method from class A was never executed.
A.doInitialization() -> THIS WAS NEVER EXECUTED silently
B.doSpecificInitialization()
B.doTests()
Playing with privacy modifiers I found that TestNG will not execute a #BeforeClass annotated method from inherited class if a method is not visible from a class-inheritor
So this will work:
public class A {
#BeforeClass
private void doInitialization() {...}
}
public class B extends A {
#BeforeClass
//Here a privacy modifier matters -> please make sure your method is public or protected so it will be visible for ancestors
protected void doSpecificInitialization() {...}
#Test
public void doTests() {...}
}
As a result following happens:
A.doInitialization()
B.doSpecificInitialization()
B.doTests()
In my case (JUnit) I have the same methods called setup() in the base class and the derived class. In this case only the derived class's method is called, and I have it call the base class method.
A better and cleaner way to achieve this using inheritance may be as following -
abstract class A {
#BeforeClass
void doInitialization() {}
}
class B extends A {
#Override
#BeforeClass
void doInitialization() {
super.doInitialization();
}
#Test
void doTests() {}
}