请注意，测试函数现在用的是 `async def`，而不是像以前使用 `TestClient` 时那样只是 `def` 。

///

我们现在可以使用应用程序创建一个 `AsyncClient` ，并使用 `await` 向其发送异步请求。

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

这相当于：

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

我们曾经通过它向 `TestClient` 发出请求。

/// tip

请注意，我们正在将 async/await 与新的 `AsyncClient` 一起使用——请求是异步的。

///

/// warning

	readHash := crc32.NewIEEE()
	wroteHash := crc32.NewIEEE()
	read := io.Writer(readHash)
	wrote := io.Writer(wroteHash)
	if debug {
		read = io.MultiWriter(read, &readBuf)
		wrote = io.MultiWriter(wrote, &wroteBuf)
	}
	debugln := func(args ...any) {
		if debug {
			fmt.Println(args...)
		}
	}
	// Inject random reader/writer sleeps.
	const maxSleep = int(1 * time.Millisecond)
	doSleep := !testing.Short()

* `seek(offset)`: Va a la posición de bytes `offset` (`int`) en el archivo.
    * Por ejemplo, `await myfile.seek(0)` iría al inicio del archivo.
    * Esto es especialmente útil si ejecutas `await myfile.read()` una vez y luego necesitas leer el contenido nuevamente.
* `close()`: Cierra el archivo.

Como todos estos métodos son métodos `async`, necesitas "await" para ellos.

      throw RuntimeException("boom!")
    }

    queue.schedule("task", TimeUnit.MILLISECONDS.toNanos(200)) {
      log.put("normal task running")
      return@schedule -1L
    }

    queue.idleLatch().await(500, TimeUnit.MILLISECONDS)

    assertThat(log.take()).isEqualTo("failing task running")
    assertThat(log.take()).isEqualTo("uncaught exception: java.lang.RuntimeException: boom!")

            try {
                try (Writer writer = new BufferedWriter(new OutputStreamWriter(Files.newOutputStream(tempFile), getCsvEncoding()))) {
                    elevateWordService.exportCsv(writer);
                } catch (final Exception e) {
                    logger.warn("Failed to process a request.", e);

import time

from fastapi import FastAPI, Request

app = FastAPI()


@app.middleware("http")
async def add_process_time_header(request: Request, call_next):
    start_time = time.perf_counter()
    response = await call_next(request)
    process_time = time.perf_counter() - start_time
    response.headers["X-Process-Time"] = str(process_time)

 *  Fix a bug in `Content-Length` reporting for gzipped streams in the Apache
    HTTP client adapter. (thanks kwuollett)
 *  Work around the Alcatel `getByInetAddress` bug (thanks k.kocel)
 *  Be more aggressive about testing pooled sockets before reuse. (thanks
    warpspin)
 *  Include `Content-Type` and `Content-Encoding` in the Apache HTTP client
    adapter. (thanks kwuollett)
 *  Add a media type class to OkHttp.
 *  Change custom header prefix:

 * either express or implied. See the License for the specific language
 * governing permissions and limitations under the License.
 */
package org.codelibs.core.log;

import java.util.logging.Level;

/**
 * Adapter that uses java.util.logging loggers.
 *
 * @author koichik
 */
class JulLoggerAdapter implements LoggerAdapter {

    protected final String sourceClass;

    protected final java.util.logging.Logger logger;

                        },
                    }
                }
            },
            "required": True,
        },
    },
)
async def create_item(request: Request):
    raw_body = await request.body()
    data = magic_data_reader(raw_body)

# Nebenläufigkeit und async / await { #concurrency-and-async-await }

Details zur `async def`-Syntax für *Pfadoperation-Funktionen* und Hintergrundinformationen zu asynchronem Code, Nebenläufigkeit und Parallelität.

## In Eile? { #in-a-hurry }

<abbr title="too long; didn't read – Zu lang; nicht gelesen"><strong>TL;DR:</strong></abbr>

Wenn Sie Bibliotheken von Dritten verwenden, die mit `await` aufgerufen werden müssen, wie zum Beispiel:

```Python