* implement {@link #load load} instead.
   *
   * <p>The returned object is serializable if {@code supplier} is serializable.
   *
   * @param supplier the supplier to be used for loading values; must never return {@code null}
   * @return a cache loader that loads values by calling {@link Supplier#get}, irrespective of the
   *     key
   */

   * {@code key}, simply waits for that thread to finish and returns its loaded value. Note that
   * multiple threads can concurrently load values for distinct keys.
   *
   * <p>Caches loaded by a {@link CacheLoader} will call {@link CacheLoader#load} to load new values
   * into the cache. Newly loaded values are added to the cache using {@code

  private static final double DESIRED_LOAD_FACTOR = 0.5;

  /**
   * Returns an array size suitable for the backing array of a hash table that uses open addressing
   * with linear probing in its implementation. The returned size is the smallest power of two that
   * can hold setSize elements with the desired load factor.
   */
  @VisibleForTesting
  static int chooseTableSize(int setSize) {

      case UNORDERED:
        return new UndirectedGraphConnections<>(
            new HashMap<N, V>(INNER_CAPACITY, INNER_LOAD_FACTOR));
      case STABLE:
        return new UndirectedGraphConnections<>(
            new LinkedHashMap<N, V>(INNER_CAPACITY, INNER_LOAD_FACTOR));
      default:
        throw new AssertionError(incidentEdgeOrder.type());
    }
  }

  private Class<?> settableFutureClass;
  private Class<?> abstractFutureClass;

  @Override
  protected void setUp() throws Exception {
    // Load the "normal" copy of SettableFuture and related classes.
    SettableFuture<?> unused = SettableFuture.create();
    // Hack to load AbstractFuture et. al. in a new classloader so that it re-reads the cancellation

 *   <li>SafeAtomicHelper: uses Atomic FieldsUpdaters to implement synchronization
 *   <li>SynchronizedHelper: uses {@code synchronized} blocks for synchronization
 * </ul>
 *
 * To force selection of our fallback strategies we load {@link AggregateFutureState} (and all of
 * {@code com.google.common.util.concurrent}) in degenerate class loaders which make certain
 * platform classes unavailable. Then we construct a test suite so we can run the normal FuturesTest

   *
   * <ul>
   *   <li>{@linkplain LoadingCache#get(Object) awaiting the result of a pending load} rather than
   *       starting a redundant one
   *   <li>eliminating the error-prone caching boilerplate
   *   <li>tracking load {@linkplain #stats statistics}
   * </ul>
   *
   * <p>Among the further improvements that {@code LoadingCache} can provide but this method cannot:
   *

    final AtomicInteger loadAllCount = new AtomicInteger();

    CacheLoader<Object, Object> baseLoader =
        new CacheLoader<Object, Object>() {
          @Override
          public Object load(Object key) {
            loadCount.incrementAndGet();
            return new Object();
          }

          @Override
          public ListenableFuture<Object> reload(Object key, Object oldValue) {

   * implement {@link #load load} instead.
   *
   * <p>The returned object is serializable if {@code supplier} is serializable.
   *
   * @param supplier the supplier to be used for loading values; must never return {@code null}
   * @return a cache loader that loads values by calling {@link Supplier#get}, irrespective of the
   *     key
   */

   * {@link CacheLoader#load} calls to {@code loader}.
   */
  static <K, V> CacheLoader<K, V> bulkLoader(final CacheLoader<K, V> loader) {
    checkNotNull(loader);
    return new CacheLoader<K, V>() {
      @Override
      public V load(K key) throws Exception {
        return loader.load(key);
      }

      @Override