* are identical, even if the input set uses a different concept of equivalence.
   *
   * <p><i>Performance notes:</i> while the power set of a set with size {@code n} is of size {@code
   * 2^n}, its memory usage is only {@code O(n)}. When the power set is constructed, the input set
   * is merely copied. Only as the power set is iterated are the individual subsets created, and

   * if the primary input succeeds, it is never invoked. If, during the invocation of {@code
   * fallback}, an exception is thrown, this exception is used as the result of the output {@code
   * Future}.
   *
   * <p>Usage example:
   *
   * <pre>{@code
   * ListenableFuture<Integer> fetchCounterFuture = ...;
   *
   * // Falling back to a zero counter in case an exception happens when
   * // processing the RPC to fetch counters.

 * supplementary code points, use ICU4J UCharacter and UnicodeSet (freeze() after building). For
 * basic text processing based on UnicodeSet use the ICU4J UnicodeSetSpanner.
 *
 * <p>Example usages:
 *
 * <pre>
 *   String trimmed = {@link #whitespace() whitespace()}.{@link #trimFrom trimFrom}(userInput);
 *   if ({@link #ascii() ascii()}.{@link #matchesAllOf matchesAllOf}(s)) { ... }</pre>
 *

 * </ul>
 *
 * <p>These features are all optional; caches can be created using all or none of them. By default,
 * cache instances created by {@code CacheBuilder} will not perform any type of eviction.
 *
 * <p>Usage example:
 *
 * <pre>{@code
 * LoadingCache<Key, Graph> graphs = CacheBuilder.newBuilder()
 *     .maximumSize(10000)
 *     .expireAfterWrite(Duration.ofMinutes(10))
 *     .removalListener(MY_LISTENER)

 * supplementary code points, use ICU4J UCharacter and UnicodeSet (freeze() after building). For
 * basic text processing based on UnicodeSet use the ICU4J UnicodeSetSpanner.
 *
 * <p>Example usages:
 *
 * <pre>
 *   String trimmed = {@link #whitespace() whitespace()}.{@link #trimFrom trimFrom}(userInput);
 *   if ({@link #ascii() ascii()}.{@link #matchesAllOf matchesAllOf}(s)) { ... }</pre>
 *