* possibly run in another thread.
 *
 * <p>{@code UncheckedExecutionException} is intended as an alternative to {@code
 * ExecutionException} when the exception thrown by a task is an unchecked exception. However, it
 * may also wrap a checked exception in some cases.
 *
 * <p>When wrapping an {@code Error} from another thread, prefer {@link ExecutionError}. When
 * wrapping a checked exception, prefer {@code ExecutionException}.
 *

   * containers with a known order other than insertion order must override this method.
   *
   * <p>Note: This default implementation is overkill (but valid) for an unordered container. An
   * equally valid implementation for an unordered container is to throw an exception. The chosen
   * implementation, however, has the advantage of working for insertion-ordered containers, as

   *
   * <p>If the graph is directed, the resultant edge will be directed; otherwise, it will be
   * undirected.
   *
   * <p>Values do not have to be unique. However, values must be non-null.
   *
   * <p>If {@code nodeU} and {@code nodeV} are not already present in this graph, this method will
   * silently {@link #addNode(Object) add} {@code nodeU} and {@code nodeV} to the graph.
   *

   * containers with a known order other than insertion order must override this method.
   *
   * <p>Note: This default implementation is overkill (but valid) for an unordered container. An
   * equally valid implementation for an unordered container is to throw an exception. The chosen
   * implementation, however, has the advantage of working for insertion-ordered containers, as

    private final MutableValueGraph<N, V> mutableValueGraph;

    Builder(ValueGraphBuilder<N, V> graphBuilder) {
      // The incidentEdgeOrder for immutable graphs is always stable. However, we don't want to
      // modify this builder, so we make a copy instead.
      this.mutableValueGraph =
          graphBuilder.copy().incidentEdgeOrder(ElementOrder.<N>stable()).build();
    }

    /**

   *     of this class typically expect for instances to have a non-null cause. At the moment, you
   *     can <i>usually</i> still preserve behavior by passing an explicit {@code null} cause. Note,
   *     however, that passing an explicit {@code null} cause prevents anyone from calling {@link
   *     #initCause} later, so it is not quite equivalent to using a constructor that omits the
   *     cause.
   */
  @Deprecated

 *       cancelled, cancellation never propagates to a task that has started to run -- neither to
 *       the callable itself nor to any {@code Future} returned by an {@code AsyncCallable}.
 *       (However, cancellation can prevent an <i>unstarted</i> task from running.) Therefore, the
 *       next task will wait for any running callable (or pending {@code Future} returned by an

   * key set contains a key if and only if this multimap maps that key to at least one value.
   *
   * <p>Changes to the returned set will update the underlying multimap, and vice versa. However,
   * <i>adding</i> to the returned set is not possible.
   */
  Set<K> keySet();

  /**
   * Returns a view collection containing the key from each key-value pair in this multimap,

don't have to do anything. The specific rules are [already bundled][okhttp3_pro] into the JAR which can be
interpreted by R8 automatically.

If you, however, don't use R8 you have to apply the rules from [this file][okhttp3_pro]. You might
also need rules from [Okio][okio] which is a dependency of this library.

       *
       * My impression is that an AtomicReferenceFieldUpdater in a static field is similarly fast to
       * Unsafe on modern JVMs (if perhaps not quite as fast as VarHandle?). However, I'm not sure
       * exactly what we've benchmarked, and we certainly haven't benchmarked as far back as JDK 8.
       * (We also haven't benchmarked under Android. We continue to use UnsafeAtomicHelper there so