int sessionCount = 10;
        ExecutorService executor = Executors.newFixedThreadPool(5);
        CountDownLatch latch = new CountDownLatch(sessionCount);

        for (int i = 0; i < sessionCount; i++) {
            final int sessionIndex = i;
            executor.submit(() -> {
                try {
                    String sessionId = "concurrent-session-" + sessionIndex;

  /**
   * @since 13.0
   */
  @Override
  public final void addListener(Listener listener, Executor executor) {
    listeners.addListener(listener, executor);
  }

  @Override
  public String toString() {
    return getClass().getSimpleName() + " [" + state() + "]";
  }

  /**
   * Attempts to execute all the listeners in {@link #listeners} while not holding the {@link
   * #monitor}.
   */

 * make sense for different situations.
 *
 * <p><b>Note:</b> The dispatcher is orthogonal to the subscriber's {@code Executor}. The dispatcher
 * controls the order in which events are dispatched, while the executor controls how (i.e. on which
 * thread) the subscriber is actually called when an event is dispatched to it.
 *
 * @author Colin Decker
 */
abstract class Dispatcher {

{* ../../docs_src/background_tasks/tutorial001.py hl[1,13] *}

**FastAPI** créera l'objet de type `BackgroundTasks` pour vous et le passera comme paramètre.

## Créer une fonction de tâche

Une fonction à exécuter comme tâche d'arrière-plan est juste une fonction standard qui peut recevoir des paramètres.

Elle peut être une fonction asynchrone (`async def`) ou une fonction normale (`def`), **FastAPI** saura la gérer correctement.

            runCalled.getAndIncrement();
          }
        },
        directExecutor());
    Runnable execute =
        new Runnable() {
          @Override
          public void run() {
            list.execute();
          }
        };
    Thread thread1 = new Thread(execute);
    Thread thread2 = new Thread(execute);
    thread1.start();
    thread2.start();
    assertEquals(0, runCalled.get());

            runCalled.getAndIncrement();
          }
        },
        directExecutor());
    Runnable execute =
        new Runnable() {
          @Override
          public void run() {
            list.execute();
          }
        };
    Thread thread1 = new Thread(execute);
    Thread thread2 = new Thread(execute);
    thread1.start();
    thread2.start();
    assertEquals(0, runCalled.get());

   * Callable}.
   *
   * @param callable the callable task
   * @since 10.0
   */
  public static <V extends @Nullable Object> ListenableFutureTask<V> create(Callable<V> callable) {
    return new ListenableFutureTask<>(callable);
  }

  /**
   * Creates a {@code ListenableFutureTask} that will upon running, execute the given {@code

   * Callable}.
   *
   * @param callable the callable task
   * @since 10.0
   */
  public static <V extends @Nullable Object> ListenableFutureTask<V> create(Callable<V> callable) {
    return new ListenableFutureTask<>(callable);
  }

  /**
   * Creates a {@code ListenableFutureTask} that will upon running, execute the given {@code

El `TestClient` hace algo de magia interna para llamar a la aplicación FastAPI asíncrona en tus funciones de test `def` normales, usando pytest estándar. Pero esa magia ya no funciona cuando lo usamos dentro de funciones asíncronas. Al ejecutar nuestros tests de manera asíncrona, ya no podemos usar el `TestClient` dentro de nuestras funciones de test.

        .build();

    Request request = new Request.Builder()
        .url("ws://echo.websocket.org")
        .build();
    client.newWebSocket(request, this);

    // Trigger shutdown of the dispatcher's executor so this process exits immediately.
    client.dispatcher().executorService().shutdown();
  }

  @Override public void onOpen(WebSocket webSocket, Response response) {
    webSocket.send("Hello...");