latch.countDown();
              }
            })
        .start();

    Uninterruptibles.awaitUninterruptibly(latch);

    // See Dispatcher.LegacyAsyncDispatcher for an explanation of why there aren't really any
    // useful testable guarantees about the behavior of that dispatcher in a multithreaded
    // environment. Here we simply test that all the expected dispatches happened in some order.

## Links

-   [GitHub project](https://github.com/google/guava)
-   [Issue tracker: Report a defect or feature request](https://github.com/google/guava/issues/new)
-   [StackOverflow: Ask "how-to" and "why-didn't-it-work" questions](https://stackoverflow.com/questions/ask?tags=guava+java)
-   [guava-announce: Announcements of releases and upcoming significant changes](https://groups.google.com/group/guava-announce)

 * might run on the {@code directExecutor()} or be otherwise re-entrant (call back into your
 * object). So it is important to not call {@link #dispatch} while holding any locks. This is why
 * {@link #enqueue} and {@link #dispatch} are 2 different methods. It is expected that the decision
 * to run a particular event is made during the state change, but the decision to actually invoke

     *       object associated with our thread, because if it was the publication wouldn't have been
     *       unsafe and we'd have seen our thread as the value. This state is also why a new
     *       ThreadConfinedTaskQueue object must be created for each inline execution, because
     *       observing a null thread does not mean the object is safe to reuse.

    private final AtomicInteger countLoad = new AtomicInteger();
    private final AtomicInteger countReload = new AtomicInteger();

    @CanIgnoreReturnValue // Sure, why not?
    @Override
    public Integer load(Integer key) {
      countLoad.incrementAndGet();
      return key;
    }

    @GwtIncompatible // reload
    @Override

    // the fast path code. The safe min/max values are very unlikely to extend
    // into the range of surrogate characters, but if they do we must not test
    // any values in that range. To see why, consider the case where:
    // safeMin <= {hi,lo} <= safeMax
    // where {hi,lo} are characters forming a surrogate pair such that:
    // codePointOf(hi, lo) > safeMax

   */
  public ThreadFactory build() {
    return doBuild(this);
  }

  // Split out so that the anonymous ThreadFactory can't contain a reference back to the builder.
  // At least, I assume that's why. TODO(cpovirk): Check, and maybe add a test for this.
  @SuppressWarnings("ThreadPriorityCheck") // We only propagate user requests (which we discourage).
  private static ThreadFactory doBuild(ThreadFactoryBuilder builder) {

      setException(t);
      return;
    } finally {
      function = null;
    }

    /*
     * If set()/setValue() throws an Error, we let it propagate. Why? The most likely Error is a
     * StackOverflowError (from deep transform(..., directExecutor()) nesting), and calling
     * setException(stackOverflowError) would fail:
     *

    checkNotNull(valueComparator);
    return Comparator.<Optional<T>, @Nullable T>comparing(
        o -> orElseNull(o), Comparator.nullsLast(valueComparator));
  }

  // For discussion of why this exists, see the Android flavor.
  private static <T> @Nullable T orElseNull(Optional<T> optional) {
    return optional.orElse(null);
  }

  /**

   *     different type of exception. For background on the deprecation, read <a
   *     href="https://github.com/google/guava/wiki/Why-we-deprecated-Throwables.propagate">Why we
   *     deprecated {@code Throwables.propagate}</a>.
   */
  @CanIgnoreReturnValue
  @J2ktIncompatible
  @GwtIncompatible
  @Deprecated