* Luego devuelve la **response**.

/// note | Detalles Técnicos

Si tienes dependencias con `yield`, el código de salida se ejecutará *después* del middleware.

Si hubiera alguna tarea en segundo plano (documentada más adelante), se ejecutará *después* de todo el middleware.

///

## Crear un middleware

Para crear un middleware usas el decorador `@app.middleware("http")` encima de una función.

   * calling thread. Tasks are run to completion before a {@code Future} is returned to the caller
   * (unless the executor has been shutdown).
   *
   * <p>The returned executor is backed by the executor returned by {@link
   * MoreExecutors#newDirectExecutorService} and subject to the same constraints.
   *
   * <p>Although all tasks are immediately executed in the thread that submitted the task, this

import java.lang.annotation.Retention;
import java.lang.annotation.RetentionPolicy;
import java.lang.annotation.Target;

/**
 * Marks an event subscriber method as being thread-safe. This annotation indicates that EventBus
 * may invoke the event subscriber simultaneously from multiple threads.
 *
 * <p>This does not mark the method, and so should be used in combination with {@link Subscribe}.
 *
 * @author Cliff Biffle
 * @since 10.0
 */

      // unpark even though the thread has already removed itself from the list. But even if we did
      // use a CAS, that race would still exist (it would just be ever so slightly smaller).
      Thread w = thread;
      if (w != null) {
        thread = null;
        LockSupport.unpark(w);
      }
    }
  }

  /*

     * simple approach would be for each caller CAS it from null to a Set populated with its
     * exception. But there's another race: If the first thread fails with an exception and a second
     * thread immediately fails with the same exception:
     *
     * Thread1: calls setException(), which returns true, context switch before it can CAS
     * seenExceptionsField to its exception
     *

      // unpark even though the thread has already removed itself from the list. But even if we did
      // use a CAS, that race would still exist (it would just be ever so slightly smaller).
      Thread w = thread;
      if (w != null) {
        thread = null;
        LockSupport.unpark(w);
      }
    }
  }

  /*

<img src="/img/tutorial/path-operation-configuration/image01.png">

### Tags con Enums

Si tienes una gran aplicación, podrías terminar acumulando **varias tags**, y querrías asegurarte de que siempre uses la **misma tag** para *path operations* relacionadas.

En estos casos, podría tener sentido almacenar las tags en un `Enum`.

Ele tem:

* Performance seriamente impressionante.
* Suporte a WebSocket.
* Suporte a GraphQL.
* Tarefas de processamento interno por trás dos panos.
* Eventos de inicialização e encerramento.
* Cliente de testes construído com HTTPX.
* Respostas CORS, GZip, Arquivos Estáticos, Streaming.
* Suporte para Sessão e Cookie.

Aqui está um pouco dessa história.

## Alternativas

Eu tenho criado APIs com requisitos complexos por vários anos (Aprendizado de Máquina, sistemas distribuídos, tarefas assíncronas, banco de dados NoSQL etc.), liderando vários times de desenvolvedores.

Como parte disso, eu precisava investigar, testar e usar muitas alternativas.

                } finally {
                    endLatch.countDown();
                }
            });
        }

        // Start all threads
        startLatch.countDown();

        // Wait for completion
        assertTrue(endLatch.await(30, TimeUnit.SECONDS), "All threads should complete");
        executor.shutdown();

        // Then - All operations should succeed