Reputation: 44124
I have a JUnit 4 test that loops through an array of test data:
public @Test void testAll() {
final Object[][] sets = new Object[][] {
// SET TYPE VALUE
// --- valid sets
// groups
x(s(A,1, B,1, C,1), G),
x(s(A,4, B,4, D,4), G),
x(s(A,8, B,8, D,8, C,8), G),
x(s(J,J, B,4, D,4), G, 4*3),
x(s(A,9, J,J, D,9), G, 9*3),
x(s(A,2, B,2, C,2), G),
x(s(A,4, B,4, J,J), G, 4*3),
x(s(A,4, B,4, C,4, D,4), G),
// runs
x(s(A,1, A,2, A,3), R),
x(s(B,8, B,9, B,10), R),
x(s(J,J, C,2, C,3), R, 6),
x(s(D,8, D,9, J,J, D,11), R, 38),
x(s(D,8, D,9, J,J, J,J), R, 38),
// sames
x(s(A,1, A,1), S),
x(s(B,4, B,4, B,4), S),
x(s(C,8, C,8), S),
x(s(D,3, D,3), S),
// doubt-cases, assume group (TODO: verify this is correct)
x(s(J,J, J,J, D,4), G, 4*3),
x(s(A,7, J,J, J,J), G, 7*3),
x(s(J,J, D,9, J,J), G, 9*3),
x(s(J,J, J,J, J,J), G, 1),
// --- invalid sets
x(s(B,1, A,2, A,3), I), // not same colour
x(s(D,11, D,12, J,J, J,J), I), // last joker is 14
x(s(B,1, B,1, A,1), I), // duplicate B1
x(s(A,1, A,2, A,3, A,5), I), // gap A4
x(s(J,J, A,1, J,J, B,1, C,1), I), // one J replaces D1, then nothing left to replace
x(s(A,1, A,2), I), // short
x(s(B,1), I), // shorter
x(s(A,5, A,6), I), // short
};
for (Object[] o : sets) {
TileSet s = (TileSet) o[0];
Type t = (Type) o[1];
int v = (Integer) o[2];
System.out.println(s);
assertEquals(t, s.getType());
assertEquals(v, s.getValue());
// test isValid, though it's Too Simple To Break(R)
if (t == Type.INVALID) assertFalse(s.isValid());
else assertTrue(s.isValid());
}
}
Because it's all in one test method, the whole test stops as soon as one element in the array fails. Is there a way around that, without making a method for each test item? Maybe something with reflection?
Upvotes: 22
Views: 37315
Reputation: 3383
Parameterized Tests in JUnit 5 have got easier with a better syntax, I think. You can read about here. One of the options is to declare a method to provide the list of arguments to test, such as below:
// Given
private static Stream<Arguments> shouldReturnExpectedResultWhenGivenRightInput() {
return Stream.of(
Arguments.of(3, "Fizz"),
Arguments.of(5, "Buzz"),
Arguments.of(8, "8"),
Arguments.of(15, "FizzBuzz"),
Arguments.of(19, "19"),
Arguments.of(30, "FizzBuzz"),
Arguments.of(40, "Buzz"),
Arguments.of(45, "FizzBuzz")
);
}
@ParameterizedTest
@MethodSource
void shouldReturnExpectedResultWhenGivenRightInput(Integer input, String expected) {
// When
val result = library.processInput(input);
// Then
assertEquals(expected, result);
}
// Given
private static Stream<Arguments> shouldNotReturnExpectedResultWhenGivenWrongInput() {
return Stream.of(
Arguments.of(3, "Buzz"),
Arguments.of(5, "Fizz"),
Arguments.of(8, "Buzz"),
Arguments.of(15, "15"),
Arguments.of(19, "Fizz")
);
}
@ParameterizedTest
@MethodSource
void shouldNotReturnExpectedResultWhenGivenWrongInput(Integer input, String unexpected) {
// When
val result = library.processInput(input);
// Then
assertNotEquals(unexpected, result);
}
Upvotes: 3
Reputation: 140051
Use JUnit 4's parameterized tests. They are a perfect fit for this type of problem, although the documentation is quite lacking.
Here are a few other samples on how to use them.:
Upvotes: 23
Reputation: 28703
catch AssertionError
and add the caught error to the errors list, at the end check the list to be empty raise a compound AssertionError
if not.
Upvotes: 5