*       and virtual memory, this is not a problem - because it is mapped read-only, the kernel
   *       can always page it to disk "for free". However, on systems where killing processes
   *       happens all the time in normal conditions (i.e., android) the OS must make a tradeoff
   *       between paging memory and killing other processes - so allocating a gigantic buffer and

    elements.add(null);
    collection = getSubjectGenerator().create(elements.toArray());
    assertEquals(
        "A Set's hashCode() should be the sum of those of its elements (with "
            + "a null element counting as having a hash of zero).",
        expectedHashCode,
        getSet().hashCode());
  }

  /**
   * Returns the {@link Method} instances for the test methods in this class which call {@code

  private static final ImmutableSet<String> NO_PS =
      ImmutableSet.of("www", "foo.ihopethiswillneverbeapublicsuffix", "x.y.z");

  /**
   * Having a public suffix is equivalent to having a registry suffix, because all registry suffixes
   * are public suffixes, and all public suffixes have registry suffixes.
   */
  private static final ImmutableSet<String> NO_RS = NO_PS;

   */
  private static final SuccessorsFunction<Character> TWO_CYCLES_GRAPH =
      createDirectedGraph("ab", "ac", "bc", "cd", "da");

  /**
   * A tree-shaped graph that looks as follows (all edges are directed facing downwards):
   *
   * <pre>{@code
   *        h
   *       /|\
   *      / | \
   *     /  |  \
   *    d   e   g
   *   /|\      |
   *  / | \     |
   * a  b  c    f
   * }</pre>

 * of the <i>next</i> request. I.e., if an expensive task arrives at an idle RateLimiter, it will be
 * granted immediately, but it is the <i>next</i> request that will experience extra throttling,
 * thus paying for the cost of the expensive task.
 *
 * @author Dimitris Andreou
 * @since 13.0
 */
// TODO(user): switch to nano precision. A natural unit of cost is "bytes", and a micro precision

   * @since 12.0 (present as {@code compare} since 2.0)
   */
  public abstract ComparisonChain compareFalseFirst(boolean left, boolean right);

  /**
   * Ends this comparison chain and returns its result: a value having the same sign as the first
   * nonzero comparison result in the chain, or zero if every result was zero.
   */
  public abstract int result();

      if (next >= 0 && next < buckets) {
        candidate = next;
      } else {
        return candidate;
      }
    }
  }

  /**
   * Returns a hash code, having the same bit length as each of the input hash codes, that combines
   * the information of these hash codes in an ordered fashion. That is, whenever two equal hash

   * objects per entry.  Instead we use an open-addressed hash table.  This design is basically
   * equivalent to RegularImmutableSet, save that instead of having a hash table containing the
   * elements directly and null for empty positions, we store indices of the keys in the hash table,
   * and ABSENT for empty positions.  We then look up the keys in alternatingKeysAndValues.
   *

        // condition efficiently without _introducing_ deadlock is probably
        // tricky. For now, just accept the race condition---missing a warning
        // now and then is still better than having no deadlock detection.
        allowedPriorLocks.put(acquiredLock, new ExampleStackTrace(acquiredLock, this));
      } else {
        // Unsafe acquisition order detected. Create and cache a

 *
 * <pre>{@code
 * void exampleBadCaller() {
 *   double d = sqrt(-1.0);
 * }
 * }</pre>
 *
 * <p>would be flagged as having called {@code sqrt()} with an illegal argument.
 *
 * <h3>Performance</h3>
 *
 * <p>Avoid passing message arguments that are expensive to compute; your code will always compute
 * them, even though they usually won't be needed. If you have such arguments, use the conventional