t: Throwable,
          response: Response?,
        ) {
          t.printStackTrace(System.out)
          latch.countDown()
        }
      },
    )

    check(latch.await(30, TimeUnit.SECONDS)) { "Timed out waiting for test $number to finish." }
    val endNanos = System.nanoTime()
    val tookMs = TimeUnit.NANOSECONDS.toMillis(endNanos - startNanos.get())
    println("Took ${tookMs}ms")
  }

    assertEquals(3L, countDownLatch.getCount());

    list.execute();

    // Verify that all of the runnables execute in a reasonable amount of time.
    assertTrue(countDownLatch.await(1L, SECONDS));
  }

  public void testExecute_idempotent() {
    AtomicInteger runCalled = new AtomicInteger();
    list.add(
        new Runnable() {
          @Override
          public void run() {

# 並行與 async / await

有關*路徑操作函式*的 `async def` 語法的細節與非同步 (asynchronous) 程式碼、並行 (concurrency) 與平行 (parallelism) 的一些背景知識。

## 趕時間嗎?

<abbr title="too long; didn't read（文長慎入）"><strong>TL;DR:</strong></abbr>

如果你正在使用要求你以 `await` 語法呼叫的第三方函式庫，例如：

```Python
results = await some_library()
```

然後，使用 `async def` 宣告你的*路徑操作函式*：


```Python hl_lines="2"
@app.get('/')
async def read_results():
    results = await some_library()

        });
        th.start();

        final CountDownLatch latch = new CountDownLatch(1);
        deferred.promise().then(response -> latch.countDown());
        assertTrue(latch.await(10, TimeUnit.SECONDS));
    }

    @Test
    public void test_doneAfterResolve() throws Exception {
        final Deferred<SuggestResponse> deferred = new Deferred<>();

        Thread th = new Thread(() -> {

* `seek(offset)`: Переміщується до позиції `offset` (`int`) у файлі.
    * Наприклад, `await myfile.seek(0)` поверне курсор на початок файлу.
    * This is especially useful if you run `await myfile.read()` once and then need to read the contents again. Це особливо корисно, якщо Ви виконуєте await `await myfile.read()` один раз, а потім потрібно знову прочитати вміст.
* `close()`: Закриває файл.

        return inner

    async def __call__(self, scope, receive, send):
        if scope["type"] != "http" or self.max_content_size is None:
            await self.app(scope, receive, send)
            return

        wrapper = self.receive_wrapper(receive)
        await self.app(scope, wrapper, send)


@router.post("/middleware")
def run_middleware(file: UploadFile = File(..., description="Big File")):

Then we can create an `AsyncClient` with the app, and send async requests to it, using `await`.

{* ../../docs_src/async_tests/app_a_py39/test_main.py hl[9:12] *}

This is the equivalent to:

```Python
response = client.get('/')
```

...that we used to make our requests with the `TestClient`.

/// tip

Note that we're using async/await with the new `AsyncClient` - the request is asynchronous.

///

/// warning

**FastAPI** provides the same `WebSocket` directly just as a convenience for you, the developer. But it comes directly from Starlette.

///

## Await for messages and send messages { #await-for-messages-and-send-messages }

In your WebSocket route you can `await` for messages and send messages.

{* ../../docs_src/websockets/tutorial001_py39.py hl[48:52] *}

You can receive and send binary, text, and JSON data.

# 并发 async / await

有关路径操作函数的 `async def` 语法以及异步代码、并发和并行的一些背景知识。

## 赶时间吗？

<abbr title="too long; didn't read(长文警告)"><strong>TL;DR:</strong></abbr>

如果你正在使用第三方库，它们会告诉你使用 `await` 关键字来调用它们，就像这样：

```Python
results = await some_library()
```

然后，通过 `async def` 声明你的 *路径操作函数*：

```Python hl_lines="2"
@app.get('/')
async def read_results():
    results = await some_library()
    return results
```

/// note

{* ../../docs_src/async_tests/app_a_py39/test_main.py hl[9:12] *}

Isso é equivalente a:

```Python
response = client.get('/')
```

...que nós utilizamos para fazer as nossas requisições utilizando o `TestClient`.

/// tip | Dica

Note que nós estamos utilizando async/await com o novo `AsyncClient` - a requisição é assíncrona.

///