* called on the filtered collection, only elements that satisfy the filter will be removed from
   * the underlying collection.
   *
   * <p>The returned collection isn't threadsafe or serializable, even if {@code unfiltered} is.
   *
   * <p>Many of the filtered collection's methods, such as {@code size()}, iterate across every

 * the context of {@code class MyStringList implements List<String>}). In such case, prefer to use
 * {@link TypeToken#resolveType} since it's simpler and more type safe. This class should only be
 * used when the type mapping isn't implied by the static type hierarchy, but provided through other
 * means such as an annotation or external configuration file.
 *
 * @author Ben Yu
 * @since 15.0
 */
@ElementTypesAreNonnullByDefault

    assertEquals(
        "Shouldn't catch any more events when unregistered.", expectedEvents, catcher2.getEvents());
  }

  // NOTE: This test will always pass if register() is thread-safe but may also
  // pass if it isn't, though this is unlikely.
  public void testRegisterThreadSafety() throws Exception {
    List<StringCatcher> catchers = Lists.newCopyOnWriteArrayList();
    List<Future<?>> futures = Lists.newArrayList();

    assertEquals(
        "Shouldn't catch any more events when unregistered.", expectedEvents, catcher2.getEvents());
  }

  // NOTE: This test will always pass if register() is thread-safe but may also
  // pass if it isn't, though this is unlikely.
  public void testRegisterThreadSafety() throws Exception {
    List<StringCatcher> catchers = Lists.newCopyOnWriteArrayList();
    List<Future<?>> futures = Lists.newArrayList();

    }
  }

  /**
   * Returns a hash function implementing FarmHash's Fingerprint64, an open-source algorithm.
   *
   * <p>This is designed for generating persistent fingerprints of strings. It isn't
   * cryptographically secure, but it produces a high-quality hash with fewer collisions than some
   * alternatives we've used in the past.
   *

   * PermittedMetaException}.
   */
  protected final class MultiExceptionListIterator implements ListIterator<E> {
    // TODO: track seen elements when order isn't guaranteed
    // TODO: verify contents afterward
    // TODO: something shiny and new instead of Stack
    // TODO: test whether null is supported (create a Feature)
    /**

  `It must have been that,' said the King, `unless it was
written to nobody, which isn't usual, you know.'

  `Who is it directed to?' said one of the jurymen.

  `It isn't directed at all,' said the White Rabbit; `in fact,
there's nothing written on the OUTSIDE.'  He unfolded the paper
as he spoke, and added `It isn't a letter, after all:  it's a set
of verses.'

          entry = newValueIndex;
        }
      } else {
        throw new IllegalArgumentException("Value already present in map: " + newValue);
      }
    }
    // we do *not* update insertion order, and it isn't a structural modification!
    deleteFromTableVToK(entry, Hashing.smearedHash(values[entry]));
    values[entry] = newValue;
    insertIntoTableVToK(entry, newValueHash);
  }

  /**

   * clear()} are called on the filtered multiset, only elements that satisfy the filter will be
   * removed from the underlying multiset.
   *
   * <p>The returned multiset isn't threadsafe or serializable, even if {@code unfiltered} is.
   *
   * <p>Many of the filtered multiset's methods, such as {@code size()}, iterate across every

    /*
     * The reason we require both arguments to be >= 0 is because otherwise, what do you return on
     * gcd(0, Long.MIN_VALUE)? BigInteger.gcd would return positive 2^63, but positive 2^63 isn't an
     * int.
     */
    checkNonNegative("a", a);
    checkNonNegative("b", b);
    if (a == 0) {
      // 0 % b == 0, so b divides a, but the converse doesn't hold.