private static final int BYTES = Stats.BYTES * 2 + Double.SIZE / Byte.SIZE;

  /**
   * Gets a byte array representation of this instance.
   *
   * <p><b>Note:</b> No guarantees are made regarding stability of the representation between
   * versions.
   */
  public byte[] toByteArray() {
    ByteBuffer buffer = ByteBuffer.allocate(BYTES).order(ByteOrder.LITTLE_ENDIAN);
    xStats.writeTo(buffer);

    MoreFiles.touch(temp);
    assertThat(Files.getLastModifiedTime(temp).toMillis()).isNotEqualTo(0);
  }

  public void testCreateParentDirectories_root() throws IOException {
    // We use a fake filesystem to sidestep flaky problems with Windows (b/136041958).
    try (FileSystem fs = Jimfs.newFileSystem(Configuration.unix())) {
      Path root = fs.getRootDirectories().iterator().next();

   *     implementation
   * @param occurrences the number of occurrences of the element to add. May be zero, in which case
   *     no change will be made.
   * @return the count of the element before the operation; possibly zero
   * @throws IllegalArgumentException if {@code occurrences} is negative, or if this operation would

 * recognized public suffix; see {@link InternetDomainName#isPublicSuffix()} for details.
 *
 * <p>Note that no network lookups are performed by any {@code HostSpecifier} methods. No attempt is
 * made to verify that a provided specifier corresponds to a real or accessible host. Only syntactic
 * and pattern-based checks are performed.
 *

    return result;
  }

  /*
   * Memory visibility of these fields. There are two cases to consider.
   *
   * 1. visibility of the writes to these fields to Fire.run:
   *
   * The initial write to delegateRef is made definitely visible via the semantics of
   * addListener/SES.schedule. The later racy write in cancel() is not guaranteed to be observed,
   * however that is fine since the correctness is based on the atomic state in our base class. The

                        "While this /may/ work fine, please look for plugin updates and/or "
                                + "request plugins be made thread-safe.",
                        "If reporting an issue, report it against the plugin in question, not against Apache Maven.")) {
                    logger.warn(s);
                }

   * this precaution.
   *
   * In part because Cells are relatively large, we avoid creating
   * them until they are needed.  When there is no contention, all
   * updates are made to the base field.  Upon first contention (a
   * failed CAS on base update), the table is initialized to size 2.
   * The table size is doubled upon further contention until

   * the {@code shutdownNow} method. First, "best-effort" with regards to canceling running tasks is
   * implemented as "no-effort". No interrupts or other attempts are made to stop threads executing
   * tasks. Second, the returned list will always be empty, as any submitted task is considered to
   * have started execution. This applies also to tasks given to {@code invokeAll} or {@code

            int behavior = random.nextInt(4);
            if (behavior == 0) { // throw an exception
              exceptionCount.incrementAndGet();
              throw new RuntimeException("fake exception for test");
            } else if (behavior == 1) { // return null
              computeNullCount.incrementAndGet();
              return null;
            } else if (behavior == 2) { // slight delay before returning

   * the iterator is created, in any way except through the iterator's own remove method, the
   * iterator will generally throw a {@link ConcurrentModificationException}. Thus, in the face of
   * concurrent modification, the iterator fails quickly and cleanly, rather than risking arbitrary,
   * non-deterministic behavior at an undetermined time in the future.
   *