* gigantic heap, in which case we scale by heap size.
   */
  private static long timeoutSeconds() {
    // This class can make no hard guarantees.  The methods in this class are inherently flaky, but
    // we try hard to make them robust in practice.  We could additionally try to add in a system
    // load timeout multiplier.  Or we could try to use a CPU time bound instead of wall clock time

    val server = serverListener.assertOpen()

    // Initiate a close on the client, which will schedule a hard cancel in 500 ms.
    val closeAtNanos = System.nanoTime()
    webSocket.close(1000, "goodbye", 500L)
    serverListener.assertClosing(1000, "goodbye")

    // Confirm that the hard cancel occurred after 500 ms.
    clientListener.assertFailure()
    val elapsedUntilFailure = System.nanoTime() - closeAtNanos

   * gigantic heap, in which case we scale by heap size.
   */
  private static long timeoutSeconds() {
    // This class can make no hard guarantees.  The methods in this class are inherently flaky, but
    // we try hard to make them robust in practice.  We could additionally try to add in a system
    // load timeout multiplier.  Or we could try to use a CPU time bound instead of wall clock time

   */
  /*
   * The new array contains nulls, even if the old array did not. If we wanted to be accurate, we
   * would declare a return type of `@Nullable T[]`. However, we've decided not to think too hard
   * about arrays for now, as they're a mess. (We previously discussed this in the review of
   * ObjectArrays, which is the main caller of this method.)
   */

   * world. Figure out what sort of space-time tradeoff we're actually going to get here with the
   * *Map variants. This class is particularly hard to benchmark, because the benefit is not only in
   * less allocation, but also having the GC do less work to scan the heap because of fewer
   * references, which is particularly hard to quantify.
   */

  /** Creates an empty {@code CompactHashMap} instance. */

   */
  /*
   * The new array contains nulls, even if the old array did not. If we wanted to be accurate, we
   * would declare a return type of `@Nullable T[]`. However, we've decided not to think too hard
   * about arrays for now, as they're a mess. (We previously discussed this in the review of
   * ObjectArrays, which is the main caller of this method.)
   */

    assertNotWellFormed(0xF0, 0xA4, 0xAD, 0xC0);
    // Special cases for byte2
    assertNotWellFormed(0xF0, 0x8F, 0xAD, 0xA2);
    assertNotWellFormed(0xF4, 0x90, 0xAD, 0xA2);
  }

  /** Tests some hard-coded test cases. */
  public void testSomeSequences() {
    // Empty
    assertWellFormed();
    // One-byte characters, including control characters

   * world. Figure out what sort of space-time tradeoff we're actually going to get here with the
   * *Map variants. This class is particularly hard to benchmark, because the benefit is not only in
   * less allocation, but also having the GC do less work to scan the heap because of fewer
   * references, which is particularly hard to quantify.
   */

  /** Creates an empty {@code CompactHashMap} instance. */

   * Function<A, B> or who call convertAll (and for any checkers that apply @PolyNull-like semantics
   * to Converter.convert). So maybe we don't want to think too hard about how to prevent our
   * checkers from issuing errors related to LegacyConverter, since it turns out that
   * LegacyConverter does violate the assumptions we make elsewhere.
   */

          // have no idea.
          // TODO(lukes): consider building everything in terms of ListenableScheduledFuture then
          // the AbstractService could monitor the future directly. Rescheduling is still hard...
          // but it would help with some of these lock ordering issues.
          scheduleFailure = e;
          toReturn = new FutureAsCancellable(immediateCancelledFuture());
        } finally {