format(
              "estimated size of spliterator after trySplit (%s) is larger than original size (%s)",
              spliterator.estimateSize(), originalSize));
    }
    if (trySplit != null) {
      if (trySplit.estimateSize() > originalSize) {
        fail(
            format(
                "estimated size of trySplit result (%s) is larger than original size (%s)",
                trySplit.estimateSize(), originalSize));

 *       amounts of computation (creating a thread, calling a few methods, etc) needed to reach a
 *       timeout point. Similarly, a SMALL is always discriminable as larger than SHORT and smaller
 *       than MEDIUM. And so on. These constants are set to conservative values, but even so, if
 *       there is ever any doubt, they can all be increased in one spot to rerun tests on slower

        ? SortedSetTestSuiteBuilder.using((TestSortedSetGenerator<K>) keySetGenerator)
        : SetTestSuiteBuilder.using(keySetGenerator);
  }

  /**
   * To avoid infinite recursion, test suites with these marker features won't have derived suites
   * created for them.
   */
  enum NoRecurse implements Feature<@Nullable Void> {
    SUBMAP,
    DESCENDING;

    @Override

    // 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.
    //
    // TODO(user): Consider using
    // java/lang/management/OperatingSystemMXBean.html#getSystemLoadAverage()

  // 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
  //   spinning, so we use SPIN_THRESHOLD_NANOS which is what AbstractQueuedSynchronizer uses for

 *       amounts of computation (creating a thread, calling a few methods, etc) needed to reach a
 *       timeout point. Similarly, a SMALL is always discriminable as larger than SHORT and smaller
 *       than MEDIUM. And so on. These constants are set to conservative values, but even so, if
 *       there is ever any doubt, they can all be increased in one spot to rerun tests on slower

     * threshold has not yet been updated, so there are no atomicity requirements for this with
     * respect to reads.
     *
     * As a guide, all critical volatile reads and writes to the count field are marked in code
     * comments.
     */

    @Weak final LocalCache<K, V> map;

    /** The number of live elements in this segment's region. */
    volatile int count;

-   [guava-discuss: For open-ended questions and discussion](https://groups.google.com/group/guava-discuss)

## IMPORTANT WARNINGS

1.  APIs marked with the `@Beta` annotation at the class or method level are
    subject to change. They can be modified in any way, or even removed, at any
    time. If your code is a library itself (i.e., it is used on the CLASSPATH of

    @Override
    @CheckForNull
    public E getNext() {
      return null;
    }
  }

  /** Marker interface for {@link InternalEntry} implementations for strong values. */
  interface StrongValueEntry<K, V, E extends InternalEntry<K, V, E>>
      extends InternalEntry<K, V, E> {}

  /** Marker interface for {@link InternalEntry} implementations for weak values. */

   * <ol>
   *   <li>execution will not stop until the task queue is empty.
   *   <li>tasks will begin execution with the thread marked as not interrupted - any interruption
   *       applies only to the task that was running at the point of interruption.
   *   <li>if the thread was interrupted before the SequentialExecutor's worker begins execution,
   *       the interrupt will be restored to the thread after it completes so that its {@code