* href="http://en.wikipedia.org/wiki/UTF-16">UTF-16</a> before calling this method.
   *
   * <p><b>Note:</b> When implementing an escaper it is a good idea to override this method for
   * efficiency by inlining the implementation of {@link #nextEscapeIndex(CharSequence, int, int)}
   * directly. Doing this for {@link com.google.common.net.PercentEscaper} more than doubled the

    Ordering<Integer> e = integers.compound(numbers);

    // This works with three levels too (IDEA falsely reports errors as noted
    // below. Both javac and eclipse handle these cases correctly.)

    Ordering<Number> f = numbers.compound(objects).compound(objects); // bad IDEA
    Ordering<Number> g = objects.compound(numbers).compound(objects);

    Ordering<Integer> e = integers.compound(numbers);

    // This works with three levels too (IDEA falsely reports errors as noted
    // below. Both javac and eclipse handle these cases correctly.)

    Ordering<Number> f = numbers.compound(objects).compound(objects); // bad IDEA
    Ordering<Number> g = objects.compound(numbers).compound(objects);

 *       check for thread assertion failures.
 *   <li>All delays and timeouts must use one of the constants {@code SHORT_DELAY_MS}, {@code
 *       SMALL_DELAY_MS}, {@code MEDIUM_DELAY_MS}, {@code LONG_DELAY_MS}. The idea here is that a
 *       SHORT is always discriminable from zero time, and always allows enough time for the small
 *       amounts of computation (creating a thread, calling a few methods, etc) needed to reach a

  //    Imagine:
  //    guard.lock();
  //    try { /* monitor locked and guard satisfied here */ }
  //    finally { guard.unlock(); }
  // Here are Justin's design notes about this:
  //
  // This idea has come up from time to time, and I think one of my
  // earlier versions of Monitor even did something like this. I ended
  // up strongly favoring the current interface.
  //

  }

  @GwtIncompatible // TODO
  static int log10Floor(long x) {
    /*
     * Based on Hacker's Delight Fig. 11-5, the two-table-lookup, branch-free implementation.
     *
     * The key idea is that based on the number of leading zeros (equivalently, floor(log2(x))), we
     * can narrow the possible floor(log10(x)) values to two. For example, if floor(log2(x)) is 6,

 * "the same," <i>unless specified otherwise</i> by the implementation.
 *
 * <p><b>Warning:</b> as with normal {@link Set}s, it is almost always a bad idea to modify an
 * element (in a way that affects its {@link Object#equals} behavior) while it is contained in a
 * multiset. Undefined behavior and bugs will result.
 *
 * <h3>Implementations</h3>
 *
 * <ul>

 *       check for thread assertion failures.
 *   <li>All delays and timeouts must use one of the constants {@code SHORT_DELAY_MS}, {@code
 *       SMALL_DELAY_MS}, {@code MEDIUM_DELAY_MS}, {@code LONG_DELAY_MS}. The idea here is that a
 *       SHORT is always discriminable from zero time, and always allows enough time for the small
 *       amounts of computation (creating a thread, calling a few methods, etc) needed to reach a

    }
  }

  private static int log10Floor(int x) {
    /*
     * Based on Hacker's Delight Fig. 11-5, the two-table-lookup, branch-free implementation.
     *
     * The key idea is that based on the number of leading zeros (equivalently, floor(log2(x))), we
     * can narrow the possible floor(log10(x)) values to two. For example, if floor(log2(x)) is 6,

          // because the service does not monitor the state of the future so if the exception is not
          // caught and forwarded to the service the task would stop executing but the service would
          // have no idea.
          // TODO(lukes): consider building everything in terms of ListenableScheduledFuture then
          // the AbstractService could monitor the future directly. Rescheduling is still hard...