|
57 | 57 | * deterministic but intentionally non-obvious.
|
58 | 58 | *
|
59 | 59 | * <p>In addition, {@code @BeforeAll} methods are in no way linked to
|
60 |
| - * {@code @AfterAll} methods, i.e. there are no guarantees with regard to their |
61 |
| - * wrapping behavior. For example, given two {@code @BeforeAll} methods |
62 |
| - * {@code createA()} and {@code createB()} as well as two {@code @AfterAll} |
| 60 | + * {@code @AfterAll} methods. Consequently, there are no guarantees with regard |
| 61 | + * to their <em>wrapping</em> behavior. For example, given two {@code @BeforeAll} |
| 62 | + * methods {@code createA()} and {@code createB()} as well as two {@code @AfterAll} |
63 | 63 | * methods {@code destroyA()} and {@code destroyB()}, the order in which the
|
64 | 64 | * {@code @BeforeAll} methods are executed (e.g. {@code createA()} before
|
65 | 65 | * {@code createB()}) does not imply any order for the seemingly corresponding
|
66 |
| - * {@code @AfterAll} methods: (e.g. {@code destroyA()} might be called before |
67 |
| - * <em>or</em> after {@code destroyB()}). Thus, the JUnit Team recommends that |
68 |
| - * developers declare at most one {@code @BeforeAll}/{@code @AfterAll} method |
69 |
| - * per test class/interface unless there are no dependencies between them. |
| 66 | + * {@code @AfterAll} methods. In other words, {@code destroyA()} might be called |
| 67 | + * before <em>or</em> after {@code destroyB()}. The JUnit Team therefore recommends |
| 68 | + * that developers declare at most one {@code @BeforeAll} method and at most one |
| 69 | + * {@code @AfterAll} method per test class or test interface unless there are no |
| 70 | + * dependencies between the {@code @BeforeAll} methods or between the |
| 71 | + * {@code @AfterAll} methods. |
70 | 72 | *
|
71 | 73 | * <h3>Composition</h3>
|
72 | 74 | *
|
|
0 commit comments