* subsequent requests being smoothly limited at the stable rate of {@code permitsPerSecond}.
   *
   * @param permitsPerSecond the rate of the returned {@code RateLimiter}, measured in how many
   *     permits become available per second
   * @throws IllegalArgumentException if {@code permitsPerSecond} is negative or zero
   */
  // TODO(user): "This is equivalent to
  // {@code createWithCapacity(permitsPerSecond, 1, TimeUnit.SECONDS)}".

   * includes exceptional exits, so all control paths involving signalling must be protected by a
   * finally block.
   *
   * Further optimizations of this algorithm become increasingly subtle. A wait that terminates
   * without the guard being satisfied (due to timeout, but not interrupt) can then immediately exit

 *       <i>removing</i> that key from the multimap.
 *   <li>The total entry count is available as {@link #size}.
 *   <li>Many complex operations become easier; for example, {@code
 *       Collections.min(multimap.values())} finds the smallest value across all keys.
 * </ul>
 *
 * <h3>Implementations</h3>
 *
 * <ul>
 *   <li>{@link ImmutableListMultimap}

   *
   * On the other hand, past underutilization could mean that "the server responsible for handling
   * the request has become less ready for future requests", i.e. its caches become stale, and
   * requests become more likely to trigger expensive operations (a more extreme case of this
   * example is when a server has just booted, and it is mostly busy with getting itself up to
   * speed).
   *

   * colliding threads.  Because search is random, and collisions
   * only become known via CAS failures, convergence can be slow,
   * and because threads are typically not bound to CPUS forever,
   * may not occur at all. However, despite these limitations,
   * observed contention rates are typically low in these cases.
   *
   * It is possible for a Cell to become unused when threads that
   * once hashed to it terminate, as well as in the case where

      await(done);
      if (predicate.isDone()) {
        return;
      }
    } while (System.nanoTime() - deadline < 0);
    throw formatRuntimeException(
        "Predicate did not become true within %d second timeout", timeoutSeconds);
  }

  /**
   * Waits until the given latch has {@linkplain CountDownLatch#countDown counted down} to zero,

 *       code, or even to discover <i>any</i> two distinct inputs that yield the same result. These
 *       are called <i>cryptographic hash functions</i>. But, whenever it is learned that either of
 *       these feats has become computationally feasible, the function is deemed "broken" and should
 *       no longer be used for secure purposes. (This is the likely eventual fate of <i>all</i>
 *       cryptographic hashes.)

   * colliding threads.  Because search is random, and collisions
   * only become known via CAS failures, convergence can be slow,
   * and because threads are typically not bound to CPUS forever,
   * may not occur at all. However, despite these limitations,
   * observed contention rates are typically low in these cases.
   *
   * It is possible for a Cell to become unused when threads that
   * once hashed to it terminate, as well as in the case where

    sepStopwatchA.set(null);
    GcFinalization.awaitClear(sepStopwatchRef.get());
    // Return a weak reference to the parallel ClassLoader. This is the reference that should
    // eventually become clear if there are no other references to the ClassLoader.
    return new WeakReference<ClassLoader>(sepLoader);
  }

  private void doTestUnloadable() throws Exception {

  /**
   * Returns the identity function.
   *
   * <p><b>Discouraged:</b> Prefer using a lambda like {@code v -> v}, which is shorter and often
   * more readable.
   */
  // implementation is "fully variant"; E has become a "pass-through" type
  @SuppressWarnings("unchecked")
  public static <E extends @Nullable Object> Function<E, E> identity() {
    return (Function<E, E>) IdentityFunction.INSTANCE;
  }

  // enum singleton pattern