Each test method and test class can be annotated as belonging to a category:
public static class SomeUITests {
@Category(UserAvailable.class)
@Test
public void askUserToPressAKey() { }
@Test
public void simulatePressingKey() { }
}
@Category(InternetConnected.class)
public static class InternetTests {
@Test
public void pingServer() { }
}
To run all of the tests in a particular category, you must currently explicitly create a custom request:
new JUnitCore().run(Request.aClass(SomeUITests.class).inCategories(UserAvailable.class));
This feature will very likely be improved before the final release of JUnit 4.5
@Before
and @After
methods are run before and after each set of attempted parameters
on a Theory
Refactoring removed duplication that used to exist in classes MethodRoadie and ClassRoadie
Exposed API ParameterSignature.getType()
JUnit is designed to efficiently capture developers' intentions about their code, and quickly check their code matches those intentions. Over the last year, we've been talking about what things developers would like to say about their code that have been difficult in the past, and how we can make them easier.
Two years ago, Joe Walnes built a new assertion mechanism on top of what was
then JMock 1. The method name was assertThat
, and the syntax looked like this:
assertThat(x, is(3));
assertThat(x, is(not(4)));
assertThat(responseString, either(containsString("color")).or(containsString("colour")));
assertThat(myList, hasItem("3"));
More generally:
assertThat([value], [matcher statement]);
Advantages of this assertion syntax include:
More readable and typeable: this syntax allows you to think in terms of subject, verb, object
(assert "x is 3") rather than assertEquals
, which uses verb, object, subject (assert "equals 3 x")
Combinations: any matcher statement s
can be negated (not(s)
), combined (either(s).or(t)
),
mapped to a collection (each(s)
), or used in custom combinations (afterFiveSeconds(s)
)
Readable failure messages. Compare
assertTrue(responseString.contains("color") || responseString.contains("colour"));
// ==> failure message:
// java.lang.AssertionError:
assertThat(responseString, anyOf(containsString("color"), containsString("colour")));
// ==> failure message:
// java.lang.AssertionError:
// Expected: (a string containing "color" or a string containing "colour")
// got: "Please choose a font"
Custom Matchers. By implementing the Matcher
interface yourself, you can get all of the
above benefits for your own custom assertions.
For a more thorough description of these points, see Joe Walnes's original post.
We have decided to include this API directly in JUnit. It's an extensible and readable syntax, and it enables new features, like assumptions and theories.
Some notes:
assertEquals
, assertTrue
, and so on.The second parameter of an assertThat
statement is a Matcher
.
We include the Matchers we want as static imports, like this:
import static org.hamcrest.CoreMatchers.is;
or:
import static org.hamcrest.CoreMatchers.*;
Manually importing Matcher
methods can be frustrating. Eclipse 3.3 includes the ability to
define
"Favorite" classes to import static methods from, which makes it easier
(Search for "Favorites" in the Preferences dialog).
We expect that support for static imports will improve in all Java IDEs in the future.
To allow compatibility with a wide variety of possible matchers, we have decided to include the classes from hamcrest-core, from the Hamcrest project. This is the first time that third-party classes have been included in JUnit.
JUnit currently ships with a few matchers, defined in
org.hamcrest.CoreMatchers
and org.junit.matchers.JUnitMatchers
.
To use many, many more, consider downloading the full hamcrest package.
JUnit contains special support for comparing string and array
values, giving specific information on how they differ. This is not
yet available using the assertThat
syntax, but we hope to bring
the two assert methods into closer alignment in future releases.
Ideally, the developer writing a test has control of all of the forces that might cause a test to fail.
If this isn't immediately possible, making dependencies explicit can often improve a design.
For example, if a test fails when run in a different locale than the developer intended,
it can be fixed by explicitly passing a locale to the domain code.
However, sometimes this is not desirable or possible.
It's good to be able to run a test against the code as it is currently written,
implicit assumptions and all, or to write a test that exposes a known bug.
For these situations, JUnit now includes the ability to express "assumptions":
import static org.junit.Assume.*
@Test public void filenameIncludesUsername() {
assumeThat(File.separatorChar, is('/'));
assertThat(new User("optimus").configFileName(), is("configfiles/optimus.cfg"));
}
@Test public void correctBehaviorWhenFilenameIsNull() {
assumeTrue(bugFixed("13356")); // bugFixed is not included in JUnit
assertThat(parse(null), is(new NullDocument()));
}
With this release, a failed assumption will lead to the test being marked as passing, regardless of what the code below the assumption may assert. In the future, this may change, and a failed assumption may lead to the test being ignored: however, third-party runners do not currently allow this option.
We have included assumeTrue
for convenience, but thanks to the
inclusion of Hamcrest, we do not need to create assumeEquals
,
assumeSame
, and other analogues to the assert*
methods. All of
those functionalities are subsumed in assumeThat
, with the appropriate
matcher.
A failing assumption in a @Before
or @BeforeClass
method will have the same effect
as a failing assumption in each @Test
method of the class.
More flexible and expressive assertions, combined with the ability to state assumptions clearly, lead to a new kind of statement of intent, which we call a "Theory". A test captures the intended behavior in one particular scenario. A theory captures some aspect of the intended behavior in possibly infinite numbers of potential scenarios. For example:
@RunWith(Theories.class)
public class UserTest {
@DataPoint public static String GOOD_USERNAME = "optimus";
@DataPoint public static String USERNAME_WITH_SLASH = "optimus/prime";
@Theory public void filenameIncludesUsername(String username) {
assumeThat(username, not(containsString("/")));
assertThat(new User(username).configFileName(), containsString(username));
}
}
This makes it clear that the user's filename should be included in the config file name, only if it doesn't contain a slash. Another test or theory might define what happens when a username does contain a slash.
UserTest
will attempt to run filenameIncludesUsername
on
every compatible DataPoint
defined in the class. If any of the
assumptions fail, the data point is silently ignored. If all of the
assumptions pass, but an assertion fails, the test fails.
The support for Theories has been absorbed from the Popper project, and more complete documentation can be found there.
Defining general statements in this way can jog the developer's memory about other potential data points and tests, also allows automated tools to search for new, unexpected data points that expose bugs.
This release contains other bug fixes and new features. Among them:
Annotated descriptions
Runner UIs, Filters, and Sorters operate on Descriptions of test methods and test classes. These Descriptions now include the annotations on the original Java source element, allowing for richer display of test results, and easier development of annotation-based filters.
Bug fix (1715326): assertEquals now compares all Numbers using their
native implementation of equals
. This assertion, which passed in
4.3, will now fail:
assertEquals(new Integer(1), new Long(1));
Non-integer Numbers (Floats, Doubles, BigDecimals, etc), which were compared incorrectly in 4.3, are now fixed.
assertEquals(long, long)
and assertEquals(double, double)
have
been re-introduced to the Assert
class, to take advantage of
Java's native widening conversions. Therefore, this still passes:
assertEquals(1, 1L);
The default runner for JUnit 4 test classes has been refactored.
The old version was named TestClassRunner
, and the new is named
JUnit4ClassRunner
. Likewise, OldTestClassRunner
is now
JUnit3ClassRunner
. The new design allows variations in running
individual test classes to be expressed with fewer custom classes.
For a good example, see the source to
org.junit.experimental.theories.Theories
.
The rules for determining which runner is applied by default to a test class have been simplified:
If the class has a @RunWith
annotation, the annotated runner
class is used.
If the class can be run with the JUnit 3 test runner (it
subclasses TestCase
, or contains a public static Test suite()
method), JUnit38ClassRunner is used.
Otherwise, JUnit4ClassRunner is used.
This default guess can always be overridden by an explicit
@RunWith(JUnit4ClassRunner.class)
or
@RunWith(JUnit38ClassRunner.class)
annotation.
The old class names TestClassRunner
and OldTestClassRunner
remain as deprecated.
Bug fix (1739095): Filters and Sorters work correctly on test
classes that contain a suite
method like:
public static junit.framework.Test suite() {
return new JUnit4TestAdapter(MyTest.class);
}
Bug fix (1745048): @After methods are now correctly called after a test method times out.