*
     * <p>Since the last element of the array is actually in the middle of the sorted structure, a
     * childless aunt node could be smaller, which would corrupt the invariant if this element
     * becomes the new parent of the aunt node. In that case, we first switch the last element with
     * its aunt node, before returning.
     */
    int swapWithConceptuallyLastElement(E actualLastElement) {

    return backingMap.get(element);
  }

  // Optional Operations - Modification Operations

  /**
   * {@inheritDoc}
   *
   * @throws IllegalArgumentException if the call would result in more than {@link
   *     Integer#MAX_VALUE} occurrences of {@code element} in this multiset.
   */
  @CanIgnoreReturnValue
  @Override
  public final int add(@ParametricNullness E element, int occurrences) {

      // It is possible due to races that we are currently in the expected state even though we
      // timed out. e.g. if we weren't event able to grab the lock within the timeout we would never
      // even check the guard. I don't think we care too much about this use case but it could lead
      // to a confusing error message.
      throw new TimeoutException("Timed out waiting for " + this + " to reach the RUNNING state.");
    }
  }

 * `server_max_window_bits`. This harms our ability to support these parameters:
 *
 *  * If `client_max_window_bits` is less than 15, OkHttp must close the web socket with code 1010.
 *    Otherwise it would compress values in a way that servers could not decompress.
 *  * If `server_max_window_bits` is less than 15, OkHttp will waste memory on an oversized buffer.
 *
 * See [RFC 7692, 7.1][rfc_7692] for details on negotiation process.
 *

  private final ElementOrder<N> nodeOrder;
  private final ElementOrder<E> edgeOrder;

  final MapIteratorCache<N, NetworkConnections<N, E>> nodeConnections;

  // We could make this a Map<E, EndpointPair<N>>. It would make incidentNodes(edge) slightly
  // faster, but also make Networks consume 5 to 20+% (increasing with average degree) more memory.

     *
     * T0 at 0 seconds
     * T1 at 1.05 seconds
     * T2 at 2 seconds
     * T3 at 3 seconds
     *
     * Due to the slight delay of T1, T2 would have to sleep till 2.05 seconds, and T3 would also
     * have to sleep till 3.05 seconds.
     */
    return create(permitsPerSecond, SleepingStopwatch.createFromSystemTimer());
  }

  @VisibleForTesting

    if (a == b) {
      return true;
    }
    if (a == null || b == null) {
      return false;
    }
    return doEquivalent(a, b);
  }

  /**
   * @since 10.0 (previously, subclasses would override equivalent())
   */
  @ForOverride
  protected abstract boolean doEquivalent(T a, T b);

  /**
   * <i>May</i> return {@code true} if {@code object} is a {@code Equivalence} that behaves

  @Override
  public final void run() {
    /*
     * Set runner thread before checking isDone(). If we were to check isDone() first, the task
     * might be cancelled before we set the runner thread. That would make it impossible to
     * interrupt, yet it will still run, since interruptTask will leave the runner value null,
     * allowing the CAS below to succeed.
     */
    Thread currentThread = Thread.currentThread();

    /*
     * requireNonNull is safe because this is called from the constructor after `futures` is set but
     * before releaseResources could be called (because we have not yet set up any of the listeners
     * that could call it, nor exposed this Future for users to call cancel() on).
     */
    requireNonNull(futures);

    // Corner case: List is empty.
    if (futures.isEmpty()) {
      handleAllCompleted();

    // This class can make no hard guarantees.  The methods in this class are inherently flaky, but
    // we try hard to make them robust in practice.  We could additionally try to add in a system
    // load timeout multiplier.  Or we could try to use a CPU time bound instead of wall clock time
    // bound.  But these ideas are harder to implement.  We do not try to detect or handle a
    // user-specified -XX:+DisableExplicitGC.