SamrPolicyHandle handle = new SamrPolicyHandle(mockHandle, server, access);
        handle.close(); // First close
        handle.close(); // Second close

        // Verify sendrecv for close was called only once
        verify(mockHandle, times(1)).sendrecv(any(MsrpcSamrCloseHandle.class));
    }

    @Test

你可以创建一个第一层依赖项（“dependable”），如下：

{* ../../docs_src/dependencies/tutorial005_an_py310.py hl[8:9] *}

这段代码声明了类型为 `str` 的可选查询参数 `q`，然后返回这个查询参数。

这个函数很简单（不过也没什么用），但却有助于让我们专注于了解子依赖项的工作方式。

## 第二层依赖项，“dependable”和“dependant” { #second-dependency-dependable-and-dependant }

接下来，创建另一个依赖项函数（一个“dependable”），并同时为它自身再声明一个依赖项（因此它同时也是一个“dependant”）：

{* ../../docs_src/dependencies/tutorial005_an_py310.py hl[13] *}

这里重点说明一下声明的参数：

        }
        return c;
    }

    /**
     * Merges two characters, handling voiced and semi-voiced sound marks.
     * @param c1 The first character.
     * @param c2 The second character.
     * @return The merged character.
     */
    public static char mergeChar(final char c1, final char c2) {
        if (c2 == 'ﾞ') {
            if ("ｶｷｸｹｺｻｼｽｾｿﾀﾁﾂﾃﾄﾊﾋﾌﾍﾎ".indexOf(c1) >= 0) {

        createTest("SomeOtherTest")
        then:
        def result = gradleRunner("test").build()
        result.output.contains("AnotherTest6 total executions: 2")
        // triggered only in the second overall run
        and: 'Tracing is provided'
        result.output.contains("""================
Test jvm exited unexpectedly.
Test jvm system exit trace (run: 1)
Gradle Test Executor 1 > AnotherTest6 > someTest

		t.Fatalf("unable initialize config file, %s", err)
	}

	initAllSubsystems(ctx)
	initConfigSubsystem(ctx, objLayer)

	globalIAMSys.Init(ctx, objLayer, globalEtcdClient, 2*time.Second)

	req, err := newTestRequest(http.MethodGet, "http://example.com:9000/bucket/object", 0, nil)
	if err != nil {
		t.Fatal(err)
	}

	if err != nil {
		t.Fatalf("unable create credential, %s", err)
	}

	globalActiveCred = creds

	globalIAMSys.Init(ctx, objLayer, globalEtcdClient, 2*time.Second)

	// List of test cases for validating http request authentication.
	testCases := []struct {
		req     *http.Request
		s3Error APIErrorCode
	}{
		// When request is unsigned, access denied is returned.

    assertThat(record.getThrown()).isEqualTo(EXCEPTION);
  }

  public void testConcurrentModification() throws Exception {
    // Tests for the absence of a bug where logging while iterating over the
    // stored log records causes a ConcurrentModificationException
    assertTrue(handler.getStoredLogRecords().isEmpty());
    ExampleClassUnderTest.foo();
    ExampleClassUnderTest.foo();
    for (LogRecord unused : handler.getStoredLogRecords()) {

    assertEquals("One event should be delivered.", 1, events.size());
    assertWithMessage("Correct string should be delivered.").that(events.get(0)).isEqualTo(EVENT);
  }

  /**
   * An {@link Executor} wanna-be that simply records the tasks it's given. Arguably the Worst
   * Executor Ever.
   *
   * @author cbiffle
   */
  public static class FakeExecutor implements Executor {
    List<Runnable> tasks = new ArrayList<>();

    @Override

  /** The first exception encountered if any threadAssertXXX method fails. */
  private final AtomicReference<Throwable> threadFailure = new AtomicReference<>(null);

  /**
   * Records an exception so that it can be rethrown later in the test harness thread, triggering a
   * test case failure. Only the first failure is recorded; subsequent calls to this method from
   * within the same test have no effect.
   */

#### Une attaque « professionnelle » { #a-professional-attack }

Bien sûr, les attaquants n'essaieraient pas tout cela à la main ; ils écriraient un programme pour le faire, avec éventuellement des milliers ou des millions de tests par seconde. Ils obtiendraient une lettre correcte supplémentaire à la fois.