*   It is similar to a Promise in JavaScript.
 *   It allows you to register callbacks that will be executed when the computation is complete,
 *   either successfully (resolve) or with an error (reject).
 * </p>
 *
 * <p>
 *   The Deferred class has a Promise inner class that allows you to register callbacks
 *   to be executed when the computation is complete.

   * The listener will run when the {@code Future}'s computation is {@linkplain Future#isDone()
   * complete} or, if the computation is already complete, immediately.
   *
   * <p>There is no guaranteed ordering of execution of listeners, but any listener added through
   * this method is guaranteed to be called once the computation is complete.
   *

 * asynchronously.
 *
 * <p>To attach to a {@link ListenableFuture} use {@link Futures#addCallback}.
 *
 * @author Anthony Zana
 * @since 10.0
 */
@GwtCompatible
public interface FutureCallback<V extends @Nullable Object> {
  /** Invoked with the result of the {@code Future} computation when it is successful. */
  void onSuccess(@ParametricNullness V result);

   * The listener will run when the {@code Future}'s computation is {@linkplain Future#isDone()
   * complete} or, if the computation is already complete, immediately.
   *
   * <p>There is no guaranteed ordering of execution of listeners, but any listener added through
   * this method is guaranteed to be called once the computation is complete.
   *

   *     ExecutionException} is thrown <a
   *     href="https://github.com/google/guava/wiki/CachesExplained#interruption">even if
   *     computation was interrupted by an {@code InterruptedException}</a>.)
   * @throws UncheckedExecutionException if an unchecked exception was thrown while loading the
   *     value

should be used with caution when working with untrusted models.

### Saved graphs and checkpoints

When loading untrusted serialized computation graphs (in form of a `GraphDef`,
`SavedModel`, or equivalent on-disk format), the set of computation primitives
available to TensorFlow is powerful enough that you should assume that the
TensorFlow process effectively executes arbitrary code.

   *
   * <p>The {@link OutputStream} should not be written to before or after the hand-off.
   */
  // TODO(user): Evaluate whether it makes sense to always piggyback the computation of a
  // HashCode on an existing OutputStream, compared to creating a separate OutputStream that could
  // be (optionally) be combined with another if needed (with something like
  // MultiplexingOutputStream).

    }

    @Override
    public List<MavenProject> getDownstreamProjects(MavenProject project, boolean transitive) {
        Key key = new Key(project, transitive, false);
        // Do not use computeIfAbsent here, as the computation is recursive
        // and this is not supported by computeIfAbsent.
        List<MavenProject> list = cache.get(key);
        if (list == null) {

import org.apache.maven.api.annotations.Nullable;

/**
 * Interface used to locate the root directory for a given project.
 *
 * The root locator is usually looked up from the plexus container.
 * One notable exception is the computation of the early {@code session.rootDirectory}
 * property which happens very early.  The implementation used in this case
 * will be discovered using the JDK service mechanism.
 *

   * expected</i>, but not absolutely required, to have the following properties:
   *
   * <ul>
   *   <li>Its execution does not cause any observable side effects.
   *   <li>The computation is <i>consistent with equals</i>; that is, {@link Objects#equals
   *       Objects.equals}{@code (a, b)} implies that {@code Objects.equals(function.apply(a),
   *       function.apply(b))}.
   * </ul>
   *