cheaper w.r.t. memory and CPU.
- Root credentials can now be changed easily. Before, a two-step process was required to
   change the cluster root credentials since they were used to en/decrypt the IAM data.
   So, both - the old and new credentials - had to be present at the same time during a rotation
   and the old credentials had to be removed once the rotation completed. This process is now gone.

  /** Return type from "contains the following elements" assertions. */
  public interface Ordered {
    /**
     * Attests that the expected values must not just be present but must be present in the order
     * they were given.
     */
    void inOrder();
  }

  @IgnoreJRERequirement // *should* be redundant with the annotation on SpliteratorTester
  private abstract static class GeneralSpliterator<E extends @Nullable Object> {

// Since this class only needs CAS on one field, we can avoid this bug by extending AtomicReference
// instead of using an AtomicReferenceFieldUpdater. This reference stores Thread instances
// and DONE/INTERRUPTED - they have a common ancestor of Runnable.
abstract class InterruptibleTask<T extends @Nullable Object>
    extends AtomicReference<@Nullable Runnable> implements Runnable {
  static {

  /**
   * Call the {@code setCount()} method under test, but do not check its return value. Callers
   * should use this method over {@link #setCountCheckReturnValue(Object, int)} when they expect
   * {@code setCount()} to throw an exception, as checking the return value could produce an
   * incorrect error message like "setCount() should return the original count" instead of the

  /** Return type from "contains the following elements" assertions. */
  public interface Ordered {
    /**
     * Attests that the expected values must not just be present but must be present in the order
     * they were given.
     */
    void inOrder();
  }

  private abstract static class GeneralSpliterator<E extends @Nullable Object> {
    final Spliterator<E> spliterator;

 *       Sometimes there are multiple tests per JSR166 method when the different "normal" behaviors
 *       differ significantly. And sometimes testcases cover multiple methods when they cannot be
 *       tested in isolation.
 *   <li>The documentation style for testcases is to provide as javadoc a simple sentence or two

   * possible with an unsafe cast which requires {@code keys} to actually be of type {@code K}.
   *
   * @since 11.0
   */
  /*
   * <? extends Object> is mostly the same as <?> to plain Java. But to nullness checkers, they
   * differ: <? extends Object> means "non-null types," while <?> means "all types."
   */
  @Override
  public ImmutableMap<K, V> getAllPresent(Iterable<? extends Object> keys) {
    Map<K, V> result = new LinkedHashMap<>();

e.g, To remove noncurrent versions of all objects keeping the most recent 5 noncurrent versions under the prefix `user-uploads/` 30 days after they become noncurrent ,

```
{
    "Rules": [
        {
            "ID": "Keep only most recent 5 noncurrent versions",
            "Status": "Enabled",
            "Filter": {

	switch e.getFunctionName() {
	case aggFnAvg, aggFnSum, aggFnMax, aggFnMin, aggFnCount:
		return e.evalAggregationNode(r, tableAlias)
	default:
		// TODO: traverse arguments and call aggregateRow on
		// them if they could be an ancestor of an
		// aggregation.
	}
	return nil
}

// getAggregate() implementation for each AST node follows. This is
// called after calling aggregateRow() on each input row, to calculate

                /* On NT without Unicode the fileNameLength
                 * includes the '\0' whereas on win98 it doesn't. I
                 * guess most clients only support non-unicode so
                 * they don't run into this.
                 */

                /* UPDATE: Maybe not! Could this be a Unicode alignment issue. I hope
                 * so. We cannot just comment out this method and use readString of