* href="http://llimllib.github.io/bloomfilter-tutorial/">tutorial</a> may help you understand how
 * they work.
 *
 * <p>The false positive probability ({@code FPP}) of a Bloom filter is defined as the probability
 * that {@linkplain #mightContain(Object)} will erroneously return {@code true} for an object that
 * has not actually been put in the {@code BloomFilter}.
 *

                  + count
                  + " trials.");
        }
      }
    }
  }

  /**
   * Test for avalanche. Avalanche means that output bits differ with roughly 1/2 probability on
   * different input keys. This test verifies that each possible 1-bit key delta achieves avalanche.
   *
   * <p>For more information: http://burtleburtle.net/bob/hash/avalanche.html
   */

    } else {
      return new BigInteger(digits, RANDOM_SOURCE).setBit(digits);
    }
  }

  /**
   * Equivalent to calling randomPositiveBigInteger(numBits) and then flipping the sign with 50%
   * probability.
   */
  static BigInteger randomNonZeroBigInteger(int numBits) {
    BigInteger result = randomPositiveBigInteger(numBits);
    return RANDOM_SOURCE.nextBoolean() ? result : result.negate();
  }

  /**

    } else {
      return new BigInteger(digits, RANDOM_SOURCE).setBit(digits);
    }
  }

  /**
   * Equivalent to calling randomPositiveBigInteger(numBits) and then flipping the sign with 50%
   * probability.
   */
  static BigInteger randomNonZeroBigInteger(int numBits) {
    BigInteger result = randomPositiveBigInteger(numBits);
    return RANDOM_SOURCE.nextBoolean() ? result : result.negate();
  }

  /**

     * determined experimentally to match our desired probability of false positives.
     */
    // NB: yes, this is surprisingly high, but that's what the experiments said was necessary
    // Raising this number slows the worst-case contains behavior, speeds up hashFloodingDetected,
    // and reduces the false-positive probability.
    static final int MAX_RUN_MULTIPLIER = 13;

    /**

   * that minimizes the need for remapping as {@code buckets} grows. That is, {@code
   * consistentHash(h, n)} equals:
   *
   * <ul>
   *   <li>{@code n - 1}, with approximate probability {@code 1/n}
   *   <li>{@code consistentHash(h, n - 1)}, otherwise (probability {@code 1 - 1/n})
   * </ul>
   *
   * <p>This method is suitable for the common use case of dividing work among buckets that meet the
   * following conditions:
   *

 * striped.get(key1) != striped.get(key2)}; the elements might nevertheless be mapped to the same
 * lock. The lower the number of stripes, the higher the probability of this happening.
 *
 * <p>There are three flavors of this class: {@code Striped<Lock>}, {@code Striped<Semaphore>}, and
 * {@code Striped<ReadWriteLock>}. For each type, two implementations are offered: {@linkplain

   * saving space.
   */
  @VisibleForTesting static final double MAX_LOAD_FACTOR = 1.2;

  /**
   * Maximum allowed false positive probability of detecting a hash flooding attack given random
   * input.
   */
  @VisibleForTesting static final double HASH_FLOODING_FPP = 0.001;

  /**

        .addEqualityGroup(ImmutableSet.of(1, 2, 1), ImmutableSet.of(2, 1, 1))
        .testEquals();
  }

  /**
   * The maximum allowed probability of falsely detecting a hash flooding attack if the input is
   * randomly generated.
   */
  private static final double HASH_FLOODING_FPP = 0.001;

              // number of expected insertions or don't run the test for so long.
              // Don't forget, the bloom filter slowly saturates over time and the
              // expected false positive probability goes up!
              assertThat(bloomFilter.expectedFpp()).isLessThan(safetyFalsePositiveRate);
            } while (stopwatch.elapsed(SECONDS) < 1);
          }
        };