*/
@FunctionalInterface
@GwtCompatible
public interface AsyncCallable<V extends @Nullable Object> {
  /**
   * Computes a result {@code Future}. The output {@code Future} need not be {@linkplain
   * Future#isDone done}, making {@code AsyncCallable} suitable for asynchronous derivations.
   *
   * <p>Throwing an exception from this method is equivalent to returning a failing {@link
   * ListenableFuture}.
   */

#### Shared reader thread

We can't rely on application threads to read data from the socket. Application threads are transient: sometimes they're reading and writing and sometimes they're off doing application-layer things. But the socket is permanent, and it needs constant attention: we dispatch all incoming frames so the connection is good-to-go when the application layer needs it.

* Ces schémas participeront à la constitution du schéma généré OpenAPI, et seront donc utilisés par les documentations automatiquement générées.

## Documentation automatique

Les schémas JSON de vos modèles seront intégrés au schéma OpenAPI global de votre application, et seront donc affichés dans la documentation interactive de l'API :

<img src="/img/tutorial/body/image01.png">

	wg.Add(2)
	go func() {
		defer wg.Done()
		srcSize, err := io.Copy(srcSha256, sobj)
		if err != nil {
			fmt.Printf("unreadable on source: %s (%s)\n", srcCtnt.Key, err)
			sourceFailed = true
		} else if srcSize != srcCtnt.Size {
			fmt.Printf("unreadable on source - size differs upon read: %s\n", srcCtnt.Key)
			sourceFailed = true
		}
	}()
	go func() {
		defer wg.Done()
		tgtSize, err := io.Copy(tgtSha256, tobj)

	disks := er.getDisks()
	var wg sync.WaitGroup
	var mu sync.Mutex
	r := rand.New(rand.NewSource(time.Now().UnixNano()))
	for _, i := range r.Perm(len(disks)) {
		wg.Add(1)
		go func() {
			defer wg.Done()
			if disks[i] == nil {
				return
			}
			di, err := disks[i].DiskInfo(context.Background(), DiskInfoOptions{})
			if err != nil || di.Healing {
				// - Do not consume disks which are not reachable

You can create production applications with **FastAPI** right now (and you have probably been doing it for some time), you just have to make sure that you use a version that works correctly with the rest of your code.

## Pin your `fastapi` version { #pin-your-fastapi-version }

Up to now, you have been declaring the parts of the request that you need with their types.

Taking data from:

* The path as parameters.
* Headers.
* Cookies.
* etc.

And by doing so, **FastAPI** is validating that data, converting it and generating documentation for your API automatically.

But there are situations where you might need to access the `Request` object directly.

  /** Invokes {@code unit.}{@link TimeUnit#sleep(long) sleep(sleepFor)} uninterruptibly. */
  @J2ktIncompatible
  @GwtIncompatible // concurrency
  @SuppressWarnings("GoodTime") // should accept a java.time.Duration
  public static void sleepUninterruptibly(long sleepFor, TimeUnit unit) {
    boolean interrupted = false;
    try {
      long remainingNanos = unit.toNanos(sleepFor);
      long end = System.nanoTime() + remainingNanos;

///

////

/// info

In Pydantic version 1 the configuration was done in an internal class `Config`, in Pydantic version 2 it's done in an attribute `model_config`. This attribute takes a `dict`, and to get autocompletion and inline errors you can import and use `SettingsConfigDict` to define that `dict`.

///

    List<Integer> list = new ArrayList<>(set);
    inputOrder.arrange(list);
    input = ImmutableList.copyOf(list);
  }

  @Benchmark
  int collections(int reps) {
    int dummy = 0;
    // Yes, this could be done more elegantly
    if (mutable) {
      for (int i = 0; i < reps; i++) {
        List<Integer> copy = new ArrayList<>(input);
        sort(copy);
        dummy += copy.get(0);
      }
    } else {