* descriptors when only the modules existence needs to be verified.</p>
     *
     * <p><b>Algorithm:</b>
     * If the given path is a directory, then there is a choice:
     * </p>
     * <ul>
     *   <li><b>Package hierarchy:</b> if a {@code module-info.class} file is found at the root,
     *       then builds a singleton map with the module name declared in that descriptor.</li>

   * 12.4.2</a>.
   *
   * <p>WARNING: Normally it's a smell if a class needs to be explicitly initialized, because static
   * state hurts system maintainability and testability. In cases when you have no choice while
   * interoperating with a legacy framework, this method helps to keep the code less ugly.
   *
   * @throws ExceptionInInitializerError if an exception is thrown during initialization of a class
   */

 */
public class SynchronizedSetTest extends TestCase {

  public static final Object MUTEX = new Integer(1); // something Serializable

  // TODO(cpovirk): Resolve difference between branches in their choice of mutex:
  // - The mainline uses `null` (even since the change in cl/99720576 was integrated).
  // - The backport continued to use MUTEX.
  public static Test suite() {
    return SetTestSuiteBuilder.using(

 * annotated with LargeTest.
 *
 * <p>Why use a custom annotation instead of {@code android.test.suitebuilder.annotation.Suppress}?
 * I'm not completely sure that this is the right choice, but it has various advantages:
 *
 * <ul>
 *   <li>An annotation named just "Suppress" might someday be treated by a non-Android tool as a

 * annotated with LargeTest.
 *
 * <p>Why use a custom annotation instead of {@code android.test.suitebuilder.annotation.Suppress}?
 * I'm not completely sure that this is the right choice, but it has various advantages:
 *
 * <ul>
 *   <li>An annotation named just "Suppress" might someday be treated by a non-Android tool as a

     * For example, an artifact of type {@value org.apache.maven.api.Type#JAR} can be placed
     * either on the class-path or on the module-path. The project needs to make a choice
     * (possibly using heuristic rules), then to add the dependency in only one of the options
     * identified by {@link PathType}.</p>
     *
     * @return file paths to place on the different tool options
     */

   * making this call. If it is not sorted, the results are undefined.
   *
   * <p>If there are elements in the list which compare as equal to the key, the choice of {@link
   * KeyPresentBehavior} decides which index is returned. If no elements compare as equal to the
   * key, the choice of {@link KeyAbsentBehavior} decides which index is returned.
   *
   * <p>This method runs in log(n) time on random-access lists, which offer near-constant-time

     * descriptors when only the modules existence needs to be verified.</p>
     *
     * <p><b>Algorithm:</b>
     * If the given path is a directory, then there is a choice:
     * </p>
     * <ul>
     *   <li><b>Package hierarchy:</b> if a {@code module-info.class} file is found at the root,
     *       then builds a singleton map with the module name declared in that descriptor.</li>

 */
public class SynchronizedSetTest extends TestCase {

  public static final Object MUTEX = new Integer(1); // something Serializable

  // TODO(cpovirk): Resolve difference between branches in their choice of mutex:
  // - The mainline uses `null` (even since the change in cl/99720576 was integrated).
  // - The backport continued to use MUTEX.
  public static Test suite() {
    return SetTestSuiteBuilder.using(

    Set<EndpointPair<Integer>> n1IncidentEdges = graph.incidentEdges(N1);
    Set<EndpointPair<Integer>> n2IncidentEdges = graph.incidentEdges(N2);
    assertThat(graphAsMutableGraph.removeNode(N1)).isTrue();

    // The choice of the size() method to call here is arbitrary.  We assume that if any of the Set
    // methods executes the validation check, they all will, and thus we only need to test one of