* is perpetually "flipped" (unencoded characters between position and limit).
   */
  private CharBuffer charBuffer;

  /**
   * byteBuffer holds encoded characters that have not yet been sent to the caller of the input
   * stream. When encoding it is "unflipped" (encoded bytes between 0 and position) and when
   * draining it is flipped (undrained bytes between position and limit).
   */

   *
   * <p>The returned iterable's iterator supports {@code remove()} if the iterator of the underlying
   * iterable supports it. Note that it is <i>not</i> possible to delete the last skipped element by
   * immediately calling {@code remove()} on that iterator, as the {@code Iterator} contract states
   * that a call to {@code remove()} before a call to {@code next()} will throw an {@link
   * IllegalStateException}.

 *       <i>not</i> the same string ({@code "1.00"}) we started with
 * </ol>
 *
 * <p>Note that it should still be the case that the round-tripped and original objects are
 * <i>similar</i>.
 *
 * <h3>Nullability</h3>
 *
 * <p>A converter always converts {@code null} to {@code null} and non-null references to non-null

     * https://stackoverflow.com/a/35754308/28465
     *
     * In that explanation's notation, our `overlap` check would be `x1 < y2 && y1 < x2`. We've
     * flipped one part of the check so that we're using "less than" in both cases (rather than a
     * mix of "less than" and "greater than"). We've also switched to "strictly less than" rather

  /**
   * Returns an encoding that behaves equivalently to this encoding, but adds a separator string
   * after every {@code n} characters. Any occurrences of any characters that occur in the separator
   * are skipped over in decoding.
   *
   * @throws IllegalArgumentException if any alphabet or padding characters appear in the separator
   *     string, or if {@code n <= 0}

            return true;
          }
        }
      }
      return false;
    }

    /**
     * Advances to the given entry. Returns true if the entry was valid, false if it should be
     * skipped.
     */
    boolean advanceTo(ReferenceEntry<K, V> entry) {
      try {
        long now = ticker.read();
        K key = entry.getKey();
        V value = getLiveValue(entry, now);

    long skipped = skipUpTo(in, n);
    if (skipped < n) {
      throw new EOFException(
          "reached end of stream after skipping " + skipped + " bytes; " + n + " bytes expected");
    }
  }

  /**
   * Discards up to {@code n} bytes of data from the input stream. This method will block until
   * either the full amount has been skipped or until the end of the stream is reached, whichever

  /** Counts the bytes in the given input stream using skip if possible. */
  private long countBySkipping(InputStream in) throws IOException {
    long count = 0;
    long skipped;
    while ((skipped = skipUpTo(in, Integer.MAX_VALUE)) > 0) {
      count += skipped;
    }
    return count;
  }

  /**
   * Copies the contents of this byte source to the given {@code OutputStream}. Does not close
   * {@code output}.
   *

    AtomicInteger numbersCloseCount = new AtomicInteger();
    Stream<Integer> numbers = Stream.of(1, 2, 3).onClose(numbersCloseCount::incrementAndGet);

    Stream<String> zipped = Streams.zip(letters, numbers, (a, b) -> a + ":" + b);

    zipped.close();

    assertThat(lettersCloseCount.get()).isEqualTo(1);
    assertThat(numbersCloseCount.get()).isEqualTo(1);
  }

  public void testZipFiniteWithInfinite() {

   * <p>The names are retrieved from the {@code nameSupplier} on the thread that is being renamed
   * right before each task is run. The renaming is best effort, if a {@link SecurityManager}
   * prevents the renaming then it will be skipped but the tasks will still execute.
   *
   * @param executor The executor to decorate
   * @param nameSupplier The source of names for each task
   */
  @J2ktIncompatible
  @GwtIncompatible // concurrency