* protect ourselves from that we want to make sure that tasks don't spin too much waiting for the
   * interrupting thread to complete the protocol.
   */
  /*
   * This test hangs (or maybe is just *very* slow) under Android.
   *
   * TODO(b/218700094): Ideally, get this to pass under Android. Failing that, convince ourselves

   * protect ourselves from that we want to make sure that tasks don't spin too much waiting for the
   * interrupting thread to complete the protocol.
   */
  /*
   * This test hangs (or maybe is just *very* slow) under Android.
   *
   * TODO(b/218700094): Ideally, get this to pass under Android. Failing that, convince ourselves

 * condition becomes true (no "signaling storms" due to use of {@link
 * java.util.concurrent.locks.Condition#signalAll Condition.signalAll}) and that no signals are lost
 * (no "hangs" due to incorrect use of {@link java.util.concurrent.locks.Condition#signal
 * Condition.signal}).
 *
 * <p>A thread is said to <i>occupy</i> a monitor if it has <i>entered</i> the monitor but not yet

   * to hang (possibly too quickly, but too-quick failures should be very unlikely, given that we
   * used to bail after 20ms during the expected-successful tests, and there we saw a failure rate
   * of ~1/5000, meaning that the other thread's get() call nearly always completes within 20ms if
   * it's going to complete at all).
   *

   * to hang (possibly too quickly, but too-quick failures should be very unlikely, given that we
   * used to bail after 20ms during the expected-successful tests, and there we saw a failure rate
   * of ~1/5000, meaning that the other thread's get() call nearly always completes within 20ms if
   * it's going to complete at all).
   *

  `What CAN all that green stuff be?' said Alice.  `And where
HAVE my shoulders got to?  And oh, my poor hands, how is it I
can't see you?'  She was moving them about as she spoke, but no
result seemed to follow, except a little shaking among the
distant green leaves.

  As there seemed to be no chance of getting her hands up to her
head, she tried to get her head down to them, and was delighted

     * this is no worse than what FutureTask does in that situation. Additionally, because the
     * Future was cancelled, its listeners have been run, so its consumers will not hang.
     *
     * Contrast this to the situation we have if setResult() throws, a situation described below.
     */
    I sourceResult;
    try {
      sourceResult = getDone(localInputFuture);

    long hangDelayMillis =
        (expectedOutcome == Outcome.HANG)
            ? EXPECTED_HANG_DELAY_MILLIS
            : UNEXPECTED_HANG_DELAY_MILLIS;
    boolean hung =
        !awaitUninterruptibly(callCompletedLatch, hangDelayMillis, TimeUnit.MILLISECONDS);
    if (hung) {
      assertEquals(expectedOutcome, Outcome.HANG);
    } else {
      assertNull(task.get(UNEXPECTED_HANG_DELAY_MILLIS, TimeUnit.MILLISECONDS));
    }
  }

   * @param terminationTimeout how long to wait for the executor to finish before terminating the
   *     JVM
   * @param timeUnit unit of time for the time parameter
   * @return an unmodifiable version of the input which will not hang the JVM
   */
  @J2ktIncompatible
  @GwtIncompatible // TODO
  @SuppressWarnings("GoodTime") // should accept a java.time.Duration
  public static ExecutorService getExitingExecutorService(

    long hangDelayMillis =
        (expectedOutcome == Outcome.HANG)
            ? EXPECTED_HANG_DELAY_MILLIS
            : UNEXPECTED_HANG_DELAY_MILLIS;
    boolean hung =
        !awaitUninterruptibly(callCompletedLatch, hangDelayMillis, TimeUnit.MILLISECONDS);
    if (hung) {
      assertEquals(expectedOutcome, Outcome.HANG);
    } else {
      assertNull(task.get(UNEXPECTED_HANG_DELAY_MILLIS, TimeUnit.MILLISECONDS));
    }
  }