* we change the rate.
   */
  public void testWeNeverGetABurstMoreThanOneSec() {
    RateLimiter limiter = RateLimiter.create(1.0, stopwatch);
    int[] rates = {1000, 1, 10, 1000000, 10, 1};
    for (int rate : rates) {
      int oneSecWorthOfWork = rate;
      stopwatch.sleepMillis(rate * 1000);
      limiter.setRate(rate);

 * <p>Rate limiters are often used to restrict the rate at which some physical or logical resource
 * is accessed. This is in contrast to {@link java.util.concurrent.Semaphore} which restricts the
 * number of concurrent accesses instead of the rate (note though that concurrency and rate are
 * closely related, e.g. see <a href="http://en.wikipedia.org/wiki/Little%27s_law">Little's
 * Law</a>).
 *

   * underlying model is a continuous function mapping storedPermits (from 0.0 to maxStoredPermits)
   * onto the 1/rate (i.e. intervals) that is effective at the given storedPermits. "storedPermits"
   * essentially measure unused time; we spend unused time buying/storing permits. Rate is
   * "permits / time", thus "1 / rate = time / permits". Thus, "1/rate" (time / permits) times

   * we change the rate.
   */
  public void testWeNeverGetABurstMoreThanOneSec() {
    RateLimiter limiter = RateLimiter.create(1.0, stopwatch);
    int[] rates = {1000, 1, 10, 1000000, 10, 1};
    for (int rate : rates) {
      int oneSecWorthOfWork = rate;
      stopwatch.sleepMillis(rate * 1000);
      limiter.setRate(rate);

        return new VerticalLinearTransformation(x1);
      } else {
        return withSlope((y2 - y1) / (x2 - x1));
      }
    }

    /**
     * Finish building an instance with the given slope, i.e. the rate of change of {@code y} with
     * respect to {@code x}. The slope must not be {@code NaN}. It may be infinite, in which case
     * the transformation is vertical. (If it is zero, the transformation is horizontal.)
     */

  int maximumSize;

  @Param("5000")
  int distinctKeys;

  @Param("4")
  int segments;

  // 1 means uniform likelihood of keys; higher means some keys are more popular
  // tweak this to control hit rate
  @Param("2.5")
  double concentration;

  Random random = new Random();

  LoadingCache<Integer, Integer> cache;

  int max;

  static AtomicLong requests = new AtomicLong(0);

  int maximumSize;

  @Param("5000")
  int distinctKeys;

  @Param("4")
  int segments;

  // 1 means uniform likelihood of keys; higher means some keys are more popular
  // tweak this to control hit rate
  @Param("2.5")
  double concentration;

  Random random = new Random();

  LoadingCache<Integer, Integer> cache;

  int max;

  static AtomicLong requests = new AtomicLong(0);

 * late. Also, all life cycle methods are executed with a lock held, so subclasses can safely modify
 * shared state without additional synchronization necessary for visibility to later executions of
 * the life cycle methods.
 *
 * <h3>Usage Example</h3>
 *
 * <p>Here is a sketch of a service which crawls a website and uses the scheduling capabilities to
 * rate limit itself.
 *
 * <pre>{@code

 *       token attempt, one measure of the quality of a hash function is <i>how well</i> it succeeds
 *       at this goal. Important note: it may be easy to achieve the theoretical minimum collision
 *       rate when using completely <i>random</i> sample input. The true test of a hash function is
 *       how it performs on representative real-world data, which tends to contain many hidden

    Closer closer = Closer.create();
    try {
      InputStream in = closer.register(openStream());
      return countBySkipping(in);
    } catch (IOException e) {
      // skip may not be supported... at any rate, try reading
    } finally {
      closer.close();
    }

    closer = Closer.create();
    try {
      InputStream in = closer.register(openStream());
      return ByteStreams.exhaust(in);