* case of (say) {@link HashSet}, which is an <i>implementation</i>, with semantics that were
 * largely defined by its supertype.
 *
 * <p>For field types and method return types, you should generally use the immutable type (such as
 * {@link ImmutableList}) instead of the general collection interface type (such as {@link List}).
 * This communicates to your callers all of the semantic guarantees listed above, which is almost
 * always very useful information.

 * case of (say) {@link HashSet}, which is an <i>implementation</i>, with semantics that were
 * largely defined by its supertype.
 *
 * <p>For field types and method return types, you should generally use the immutable type (such as
 * {@link ImmutableList}) instead of the general collection interface type (such as {@link List}).
 * This communicates to your callers all of the semantic guarantees listed above, which is almost
 * always very useful information.

    E[] array = (E[]) elements.toArray();
    /*
     * For a Set, we guess that it contains no duplicates. That's just a guess for purpose of
     * sizing; if the Set uses different equality semantics, it might contain duplicates according
     * to equals(), and we will deduplicate those properly, albeit at some cost in allocations.
     */
    int expectedSize =

   * inputs. All valid inputs must pass this regex, but it's semantically fine if not all inputs
   * that pass this regex are valid -- only a performance hit is incurred, not a semantics bug.
   */
  @GwtIncompatible // regular expressions
  static final
  java.util.regex.Pattern
      FLOATING_POINT_PATTERN = fpPattern();

  @GwtIncompatible // regular expressions
  private static

  // 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.
  // TODO(user,lukes): consider adding volatile to all these fields since in current known JVMs

 *
 * <h3>Cancelling</h3>
 *
 * Any step in a pipeline can be {@linkplain #cancel(boolean) cancelled}, even after another step
 * has been derived, with the same semantics as cancelling a {@link Future}. In addition, a
 * successfully cancelled step will immediately start closing all objects captured for later closing
 * by it and by its input steps.
 *
 * <h3>Ending a pipeline</h3>
 *

  /**
   * The HTTP <a href="http://html.spec.whatwg.org/multipage/semantics.html#hyperlink-auditing">
   * {@code Ping-From}</a> header field name.
   *
   * @since 19.0
   */
  public static final String PING_FROM = "Ping-From";
  /**
   * The HTTP <a href="http://html.spec.whatwg.org/multipage/semantics.html#hyperlink-auditing">
   * {@code Ping-To}</a> header field name.
   *
   * @since 19.0