*
   * <p>The current implementation uses the concurrency level to create a fixed number of hashtable
   * segments, each governed by its own write lock. The segment lock is taken once for each explicit
   * write, and twice for each cache loading computation (once prior to loading the new value, and
   * once after loading completes). Much internal cache management is performed at the segment

      return super.equals(obj);
    }

    @Override
    public int hashCode() {
      /*
       * The sum of the sums of the hash codes in each subset is just the sum of
       * each input element's hash code times the number of sets that element
       * appears in. Each element appears in exactly half of the 2^n sets, so:
       */
      return inputSet.keySet().hashCode() << (inputSet.size() - 1);
    }

  }

  /**
   * Returns a {@code Collector} accumulating entries into a {@code Multimap} generated from the
   * specified supplier. Each input element is mapped to a key and a stream of values, each of which
   * are put into the resulting {@code Multimap}, in the encounter order of the stream and the
   * encounter order of the streams of values.
   *
   * <p>Example:
   *

  }

  /**
   * Returns a {@code Collector} accumulating entries into a {@code Multimap} generated from the
   * specified supplier. Each input element is mapped to a key and a stream of values, each of which
   * are put into the resulting {@code Multimap}, in the encounter order of the stream and the
   * encounter order of the streams of values.
   *
   * <p>Example:
   *

   * predicate}, or {@code -1} if the Iterator has no such elements.
   *
   * <p>More formally, returns the lowest index {@code i} such that {@code
   * predicate.apply(Iterators.get(iterator, i))} returns {@code true}, or {@code -1} if there is no
   * such index.
   *
   * <p>If -1 is returned, the iterator will be left exhausted: its {@code hasNext()} method will

   * character, then iterates the remainder of the sequence calling {@link #matches(char)} for each
   * character.
   *
   * @param sequence the character sequence to replace matching characters in
   * @param replacement the character to append to the result string in place of each matching
   *     character in {@code sequence}
   * @return the new string
   */

 * them, even though they usually won't be needed. If you have such arguments, use the conventional
 * if/throw idiom instead.
 *
 * <p>Depending on your message arguments, memory may be allocated for boxing and varargs array
 * creation. However, the methods of this class have a large number of overloads that prevent such
 * allocations in many common cases.
 *

  // way in such a visibility issue would surface is most likely as a failure of cancel() to
  // propagate to the input. Cancellation propagation is fundamentally racy so this is fine.
  //
  // Future versions of the JMM may revise safe construction semantics in such a way that we can
  // safely publish these objects and we won't need this whole discussion.

 * hostname. Building APIs that consume such malformed values is difficult!
 *
 * This class has a modern API. It avoids punitive checked exceptions: [toHttpUrl] throws
 * [IllegalArgumentException] on invalid input or [toHttpUrlOrNull] returns null if the input is an
 * invalid URL. You can even be explicit about whether each component has been encoded already.
 *

   * manner.
   *
   * <p>For each of the four combinations of strong/weak key and strong/weak value, there are
   * corresponding {@link InternalEntry}, {@link Segment}, and {@link InternalEntryHelper}
   * implementations.
   *
   * @param <K> the type of the key in each entry
   * @param <V> the type of the value in each entry
   * @param <E> the type of the {@link InternalEntry} entry implementation