*
   * <pre>{@code
   * // Falling back to a zero counter in case an exception happens when processing the RPC to fetch
   * // counters.
   * ListenableFuture<Integer> faultTolerantFuture =
   *     fetchCounters().catching(FetchException.class, x -> 0, directExecutor());
   * }</pre>
   *
   * <p>When selecting an executor, note that {@code directExecutor} is dangerous in some cases. See

  }

  /**
   * Returns a hash function implementing the MD5 hash algorithm (128 hash bits).
   *
   * @deprecated If you must interoperate with a system that requires MD5, then use this method,
   *     despite its deprecation. But if you can choose your hash function, avoid MD5, which is
   *     neither fast nor secure. As of January 2017, we suggest:
   *     <ul>
   *       <li>For security:

      throw new IllegalArgumentException("key already present: " + key);
    }

    /*
     * The ordering here is important: if we deleted the key entry and then the value entry,
     * the key entry's prev or next pointer might point to the dead value entry, and when we
     * put the new entry in the key entry's position in iteration order, it might invalidate
     * the linked list.
     */

    if (oldEntryForValue != null) {

       * that behavior is undefined when the backing map is modified through another API. (It even
       * permits us to throw IllegalStateException. Maybe we should have done that, but we probably
       * shouldn't change now for fear of breaking people.)
       *
       * If the entry is still in the map, then updateIndex ensured that `index` points to the right

        "R0.50");

    limiter.setRate(Double.POSITIVE_INFINITY);
    limiter.acquire();
    limiter.acquire();
    limiter.acquire();
    assertEvents("R0.50", "R0.00", "R0.00"); // we repay the last request (.5sec), then back to +oo
  }

  /** https://code.google.com/p/guava-libraries/issues/detail?id=1791 */
  public void testInfinity_BustyTimeElapsed() {
    RateLimiter limiter = RateLimiter.create(Double.POSITIVE_INFINITY, stopwatch);

 * {@code asMap} has a few subtleties. As long as the set of keys remains unchanged, adding or
 * removing mappings does not affect the key iteration order. However, if you remove all values
 * associated with a key and then add the key back to the multimap, that key will come last in the
 * key iteration order.
 *
 * <p>The multimap does not store duplicate key-value pairs. Adding a new key-value pair equal to an

        .inOrder();
  }

  public void testCopyOf_array_empty() {
    /*
     * We don't guarantee the same-as property, so we aren't obligated to test it. However, it's
     * useful in testing - when two things are the same then one can't have bugs the other doesn't.
     */
    assertThat(ImmutableDoubleArray.copyOf(new double[0]))
        .isSameInstanceAs(ImmutableDoubleArray.of());
  }

   *
   * <p>Despite the method name, this method attempts to avoid actually copying the data when it is
   * safe to do so. The exact circumstances under which a copy will or will not be performed are
   * undocumented and subject to change.
   *
   * <p>Note that if {@code list} is a {@code List<String>}, then {@code ImmutableList.copyOf(list)}

   * multimap.asMap()} view. Repeated occurrences of an entry in the multimap after the first are
   * ignored.
   *
   * <p>Despite the method name, this method attempts to avoid actually copying the data when it is
   * safe to do so. The exact circumstances under which a copy will or will not be performed are
   * undocumented and subject to change.
   *

   *
   * <p>The resulting collection's iterator does not support {@code remove()}, but all other
   * collection methods are supported. When given an element that doesn't satisfy the predicate, the
   * collection's {@code add()} and {@code addAll()} methods throw an {@link
   * IllegalArgumentException}. When methods such as {@code removeAll()} and {@code clear()} are