* (i.e. seconds since Epoch 1970). Times should be the same as those
     * reported using the properties dialog of the Windows Explorer program.
     *
     * For Win95/98/Me this is actually the last write time. It is currently
     * not possible to retrieve the create time from files on these systems.
     *
     * @return The number of milliseconds since the 00:00:00 GMT, January 1,

  }

  /**
   * Returns the total number of times that {@link Cache} lookup methods attempted to load new
   * values. This includes both successful load operations and those that threw exceptions. This is
   * defined as {@code loadSuccessCount + loadExceptionCount}.
   *
   * <p><b>Note:</b> the values of the metrics are undefined in case of overflow (though it is

   * across calls to multiple listeners. Specifically, a given listener will have its callbacks
   * invoked in the same order as the underlying service enters those states. Additionally, at most
   * one of the listener's callbacks will execute at once. However, multiple listeners' callbacks
   * may execute concurrently, and listeners may execute in an order different from the one in which

   * works better than Ordering.natural().nullsFirst() because, if null comes before all other
   * values, it lies outside the submap/submultiset ranges we test, and the variety of tests that
   * exercise null handling fail on those subcollections.
   */
  public abstract static class NullsBefore implements Comparator<@Nullable String>, Serializable {
    private final String justAfterNull;

    protected NullsBefore(String justAfterNull) {

   * TreeMultimap#create(Comparator, Comparator)} won't suffice.
   *
   * <p>Note: the multimap assumes complete ownership over of {@code map} and the collections
   * returned by {@code factory}. Those objects should not be manually updated and they should not
   * use soft, weak, or phantom references.
   *
   * @param map place to store the mapping from each key to its corresponding values

   * TreeMultimap#create(Comparator, Comparator)} won't suffice.
   *
   * <p>Note: the multimap assumes complete ownership over of {@code map} and the collections
   * returned by {@code factory}. Those objects should not be manually updated and they should not
   * use soft, weak, or phantom references.
   *
   * @param map place to store the mapping from each key to its corresponding values

       *   either this Future was cancelled (which we ruled out with the isCancelled check above),
       *   or it had already failed. (It couldn't have completed *successfully* or even had
       *   setFuture called on it: Neither of those can happen until we've finished processing all
       *   the completed inputs. And we're still processing at least one input, the one that
       *   triggered handleException.)
       *

 * sorting with {@link java.util.Arrays#sort(double[]) Arrays.sort(double[])} or {@link
 * java.util.Collections#sort(java.util.List) Collections.sort(List<Double>)} and selecting
 * the required value(s). Those methods would sort {@link Double#NaN NaN} as if it is greater than
 * any other value and place them at the end of the dataset, even after {@link
 * Double#POSITIVE_INFINITY POSITIVE_INFINITY}.)
 *

   *
   * <p>TODO: b/147144588 - We are currently also missing the methods inherited from {@link
   * AnnotatedElement}, which {@code TypeVariable} began to extend only in Java 8. Those methods
   * refer only to types present under Android, so we could implement them in {@code
   * TypeVariableImpl} today. (We could probably then make {@code TypeVariableImpl} implement {@code

        lockStateChanged(this);
      }
    }
  }

  final class CycleDetectingReentrantReadWriteLock extends ReentrantReadWriteLock
      implements CycleDetectingLock {

    // These ReadLock/WriteLock implementations shadow those in the
    // ReentrantReadWriteLock superclass. They are simply wrappers around the
    // internal Sync object, so this is safe since the shadowed locks are never
    // exposed or used.