|| (stimuli.length >= 5 && removes > 2)) {
      // removes are the most expensive thing to test, since they often throw exceptions with stack
      // traces, so we test them a bit less aggressively
      return;
    }

    MultiExceptionListIterator reference = new MultiExceptionListIterator(expectedElements);
    I target = newTargetIterator();
    for (int i = 0; i < stimuli.length; i++) {

  public void testExpectedFpp() {
    BloomFilter<Object> bf = BloomFilter.create(HashTestUtils.BAD_FUNNEL, 10, 0.03);
    double fpp = bf.expectedFpp();
    assertThat(fpp).isEqualTo(0.0);
    // usually completed in less than 200 iterations
    while (fpp != 1.0) {
      boolean changed = bf.put(new Object());
      double newFpp = bf.expectedFpp();
      // if changed, the new fpp is strictly higher, otherwise it is the same

   * @since 20.0
   */
  public static BigInteger ceilingPowerOfTwo(BigInteger x) {
    return BigInteger.ZERO.setBit(log2(x, CEILING));
  }

  /**
   * Returns the largest power of two less than or equal to {@code x}. This is equivalent to {@code
   * BigInteger.valueOf(2).pow(log2(x, FLOOR))}.
   *
   * @throws IllegalArgumentException if {@code x <= 0}
   * @since 20.0
   */

 * NullPointerTester} uses best effort to pick non-null default values for many common JDK and Guava
 * types, and also for interfaces and public classes that have public parameter-less constructors.
 * When the non-null default value for a particular parameter type cannot be provided by {@code
 * NullPointerTester}, the caller can provide a custom non-null default value for the parameter type

    int skipped = advance(iterator, position);
    if (!iterator.hasNext()) {
      throw new IndexOutOfBoundsException(
          "position ("
              + position
              + ") must be less than the number of elements that remained ("
              + skipped
              + ")");
    }
    return iterator.next();
  }

  /**

   * resulting segment inside the cache <i>independently</i> limits its own size to approximately
   * {@code maximumSize / concurrencyLevel}.
   *
   * <p>When eviction is necessary, the cache evicts entries that are less likely to be used again.
   * For example, the cache may evict an entry because it hasn't been used recently or very often.
   *
   * <p>If {@code maximumSize} is zero, elements will be evicted immediately after being loaded into

   *   <li>the appropriate truncation indicator may be locale-dependent
   *   <li>it is safe to use non-ASCII characters in the truncation indicator
   * </ul>
   *
   * @throws IllegalArgumentException if {@code maxLength} is less than the length of {@code
   *     truncationIndicator}
   * @since 16.0
   */
  public static String truncate(CharSequence seq, int maxLength, String truncationIndicator) {
    checkNotNull(seq);

An interesting detail is that the percentage of the **CPU used** by each process can **vary** a lot over time, but the **memory (RAM)** normally stays more or less **stable**.

If you have an API that does a comparable amount of computations every time and you have a lot of clients, then the **CPU utilization** will probably *also be stable* (instead of constantly going up and down quickly).

      for (int i = 0; i < params.size(); i++) {
        TypeToken<?> paramType = type.resolveType(typeVars[i]);
        // We require all @Generates methods to either be parameter-less or accept non-null
        // values for their generic parameter types.
        Object argValue = generate(paramType);
        if (argValue == null) {
          // When a parameter of a @Generates method cannot be created,

     * Luckily, our nullness checker is smart enough to realize that `convert` has @PolyNull-like
     * behavior, so it knows that `convert(a)` returns a non-nullable value, and we don't need to
     * perform even a cast, much less a runtime check.
     *
     * All that said, don't forget that everyone should call converter.convert() instead of
     * converter.apply(), anyway. If clients use only converter.convert(), then their nullness