}
  }

  static final LazyLogger log = new LazyLogger(AbstractFuture.class);

  // A heuristic for timed gets. If the remaining timeout is less than this, spin instead of
  // blocking. This value is what AbstractQueuedSynchronizer uses.
  private static final long SPIN_THRESHOLD_NANOS = 1000L;

  private static final AtomicHelper ATOMIC_HELPER;

  static {
    AtomicHelper helper;

 *   }
 * }
 *
 * TypeToken<List<String>> stringListType = Util.<String>listType();
 * }</pre>
 *   <li>Capture a generic type with a (usually anonymous) subclass and resolve it against a context
 *       class that knows what the type parameters are. For example:
 *       <pre>{@code
 * abstract class IKnowMyType<T> {
 *   TypeToken<T> type = new TypeToken<T>(getClass()) {};
 * }
 * new IKnowMyType<String>() {}.type => String
 * }</pre>

    return set;
  }

  /**
   * Returns a new hash set using the smallest initial table size that can hold {@code expectedSize}
   * elements without resizing. Note that this is not what {@link HashSet#HashSet(int)} does, but it
   * is what most users want and expect it to do.
   *
   * <p>This behavior can't be broadly guaranteed, but has been tested with OpenJDK 1.7 and 1.8.
   *

   * they complete. Delegate futures return the same value or throw the same exception as the
   * corresponding input future returns/throws.
   *
   * <p>"In the order that they complete" means, for practical purposes, about what you would
   * expect, but there are some subtleties. First, we do guarantee that, if the output future at
   * index n is done, the output future at index n-1 is also done. (But as usual with futures, some