* both use cases by using @PolyNull. (By contrast, we can't use @PolyNull for our superinterface
   * (`implements Function<@PolyNull A, @PolyNull B>`), at least as far as I know.)
   */
  public Iterable<B> convertAll(Iterable<? extends A> fromIterable) {
    checkNotNull(fromIterable, "fromIterable");
    return () ->
        new Iterator<B>() {

      return true
    }
  }

  /**
   * HTTP/2 can have both stream timeouts (due to a problem with a single stream) and connection
   * timeouts (due to a problem with the transport). When a stream times out we don't know whether
   * the problem impacts just one stream or the entire connection.
   *
   * To differentiate the two cases we ping the server when a stream times out. If the overall

                 * *get* call in getValue but not for the *put* call here.)
                 *
                 * (Arguably we should use requireNonNull rather than uncheckedCastNullableTToT: We
                 * know that V is a non-null type because that's the only kind of value type that
                 * StandardTable supports. Thus, requireNonNull is safe as long as the cell is still

        return unsafeNull(); // See the discussion in getValue().
      }
      /*
       * The cast is safe because updateIndex found the entry for this key. (If it hadn't, then we
       * would have returned above.) Thus, we know that it and its corresponding value are in
       * position `index`.
       */
      V oldValue = uncheckedCastNullableTToT(values[index]);
      if (Objects.equals(oldValue, value)) {
        return value;
      }

   * <p>Calling {@code setDistinctValues(type, v1, v2)} also sets the default value for {@code type}
   * that's used for {@link #testNulls}.
   *
   * <p>Only necessary for types where {@link ClassSanityTester} doesn't already know how to create
   * distinct values.
   *
   * @return this tester instance
   * @since 17.0
   */
  @CanIgnoreReturnValue
  public <T> ClassSanityTester setDistinctValues(Class<T> type, T value1, T value2) {

It has an integrated dependency injection system, inspired by Angular 2. It requires pre-registering the "injectables" (like all the other dependency injection systems I know), so, it adds to the verbosity and code repetition.

As the parameters are described with TypeScript types (similar to Python type hints), editor support is quite good.

    if (!iterator.hasNext()) {
      return Optional.absent();
    }

    /*
     * TODO(kevinb): consider whether this "optimization" is worthwhile. Users with SortedSets tend
     * to know they are SortedSets and probably would not call this method.
     */
    if (iterable instanceof SortedSet) {
      SortedSet<E> sortedSet = (SortedSet<E>) iterable;
      return Optional.of(sortedSet.last());
    }

   * - The map has not switched to delegating to a java.util implementation to mitigate hash
   *   flooding. Callers can confirm this by null-checking delegateOrNull().
   *
   * In an ideal world, we would document why we know those things are true every time we call these
   * methods. But that is a bit too painful....
   */

  private Object requireTable() {
    return requireNonNull(table);
  }

    if (!iterator.hasNext()) {
      return Optional.absent();
    }

    /*
     * TODO(kevinb): consider whether this "optimization" is worthwhile. Users with SortedSets tend
     * to know they are SortedSets and probably would not call this method.
     */
    if (iterable instanceof SortedSet) {
      SortedSet<E> sortedSet = (SortedSet<E>) iterable;
      return Optional.of(sortedSet.last());
    }

     * other networking protocols the InputStream may need
     * to be passed to the readXxxWireFormat methods. This is
     * actually purer. However, in the case of smb we know the
     * wordCount and byteCount. And since every subclass of
     * ServerMessageBlock would have to perform the same read
     * operation on the input stream, we might as will pull that