*         .put(1, "one")
   *         .put(2, "two")
   *         .put(3, "three")
   *         .buildOrThrow();
   * }</pre>
   *
   * <p>For <i>small</i> immutable sorted maps, the {@code ImmutableSortedMap.of()} methods are even
   * more convenient.
   *
   * <p>Builder instances can be reused - it is safe to call {@link #buildOrThrow} multiple times to

  // Timed Get
  // There are a few design constraints to consider
  // * We want to be responsive to small timeouts, unpark() has non trivial latency overheads (I
  //   have observed 12 micros on 64-bit linux systems to wake up a parked thread). So if the
  //   timeout is small we shouldn't park(). This needs to be traded off with the cpu overhead of

   *         .put(1, "one")
   *         .put(2, "two")
   *         .put(3, "three")
   *         .buildOrThrow();
   * }</pre>
   *
   * <p>For <i>small</i> immutable sorted maps, the {@code ImmutableSortedMap.of()} methods are even
   * more convenient.
   *
   * <p>Builder instances can be reused - it is safe to call {@link #buildOrThrow} multiple times to

   * using a {@code LinkedHashSet} instead, at the cost of increased memory footprint, to get
   * deterministic iteration behavior.
   *
   * <p>This method is just a small convenience, either for {@code newHashSet(}{@link Arrays#asList
   * asList}{@code (...))}, or for creating an empty set then calling {@link Collections#addAll}.

   *
   * <p>This implementation tries to be smart in a number of ways. It recognizes cases where the
   * negation is cheaper to precompute than the matcher itself; it tries to build small hash tables
   * for matchers that only match a few characters, and so on. In the worst-case scenario, it
   * constructs an eight-kilobyte bit array and queries that. In many situations this produces a

    int threshold;

    /** The per-segment table. */
    @CheckForNull volatile AtomicReferenceArray<E> table;

    /**
     * A counter of the number of reads since the last write, used to drain queues on a small
     * fraction of read operations.
     */
    final AtomicInteger readCount = new AtomicInteger();

    Segment(MapMakerInternalMap<K, V, E, S> map, int initialCapacity) {
      this.map = map;

   *
   * <p>This implementation tries to be smart in a number of ways. It recognizes cases where the
   * negation is cheaper to precompute than the matcher itself; it tries to build small hash tables
   * for matchers that only match a few characters, and so on. In the worst-case scenario, it
   * constructs an eight-kilobyte bit array and queries that. In many situations this produces a