available, we need to hold that thread until either the bytes arrive, the stream is closed, or a timeout elapses. If we get bytes but there's nobody asking for them, we buffer them. We don't consider bytes as delivered for flow control until they're consumed by the application.

Consider an application streaming a video over http/2. Perhaps the user pauses the video and the application stops reading bytes from this stream. The buffer will fill up, and flow control prevents the server from sending...

        if (ki < key.length()) {
            key = key.substring(0, ki);
        }

        /* Subtract the server and share from the pathConsumed so that
         * it refects the part of the relative path consumed and not
         * the entire path.
         */
        dr.pathConsumed -= 1 + server.length() + 1 + share.length();

        if (referrals != null && (System.currentTimeMillis() + 10000) > referrals.expiration) {

                    dr.stripPathConsumed(1 + domain.length() + 1 + root.length());

                    if ( dr.getPathConsumed() > ( path != null ? path.length() : 0 ) ) {
                        log.error("Consumed more than we provided");
                    }

                    link = path != null && dr.getPathConsumed() > 0 ? path.substring(0, dr.getPathConsumed()) : "\\";
                    dr.setLink(link);

@app.get("/items/{item_id}")
async def read_item(item_id: int, q: Union[str, None] = None):
    return {"item_id": item_id, "q": q}
```

**Note**

Si vous n'êtes pas familier avec cette notion, consultez la section _"Vous êtes pressés ?"_ à propos de <a href="https://fastapi.tiangolo.com/fr/async/#vous-etes-presses" target="_blank">`async` et `await` dans la documentation</a>.

</details>

### Lancez

  /**
   * Null once the file has a complete copy of the upstream bytes. Only the [upstreamReader] thread
   * may access this source.
   */
  var upstream: Source?,
  /** The number of bytes consumed from [upstream]. Guarded by this. */
  var upstreamPos: Long,
  /** User-supplied additional data persisted with the source data. */
  private val metadata: ByteString,
  /** The maximum size of [buffer]. */

        tag = DerHeader.TAG_END_OF_CONTENTS,
        constructed = false,
        length = -1L,
      )
  }

  /** A source that keeps track of how many bytes it's consumed. */
  private class CountingSource(source: Source) : ForwardingSource(source) {
    var bytesRead = 0L

    override fun read(
      sink: Buffer,
      byteCount: Long,
    ): Long {

  }

  /**
   * Adds each member of {@code elements} as a candidate for the top {@code k} elements. This
   * operation takes amortized linear time in the length of {@code elements}. The iterator is
   * consumed after this operation completes.
   *
   * <p>If all input data to this {@code TopKSelector} is in a single {@code Iterator}, prefer
   * {@link Ordering#leastOf(Iterator, int)}, which provides a simpler API for that use case.

      service.shutdown();
    }
    assertThat(holder.count).isEqualTo(ITERATION_COUNT);
    for (int length : completeLengthChecks) {
      // Verify that at max depth, less than one stack frame per submitted task was consumed
      assertThat(length - baseStackDepth).isLessThan(DIRECT_EXECUTIONS_PER_THREAD / 2);
    }
  }

  @SuppressWarnings("ObjectToString") // Intended behavior
  public void testToString() {

Vous pouvez le combiner avec Gunicorn, pour avoir un serveur multi-processus asynchrone.

Pour plus de détails, consultez la section [Déploiement](deployment/index.md){.internal-link target=_blank}.

///

## Benchmarks et vitesse

* los contenidos que tu programa da al sistema para ser escritos en disco
* una operación relacionada con una API remota
* una operación de base de datos
* el retorno de resultados de una consulta de base de datos
* etc.