Usando `Depends(scope="function")`, o código de saída após o `yield` é executado logo depois que a *função de operação de rota* termina, antes de a response ser enviada de volta ao cliente.

E ao usar `Depends(scope="request")` (o padrão), o código de saída após o `yield` é executado depois que a response é enviada.

// https://docs.aws.amazon.com/AmazonS3/latest/userguide/checking-object-integrity.html
// Returns ErrInvalidChecksum if a problem with the checksum is found.
func (r *Reader) AddChecksum(req *http.Request, ignoreValue bool) error {
	cs, err := GetContentChecksum(req.Header)
	if err != nil {
		return ErrInvalidChecksum
	}
	if cs == nil {
		return nil
	}
	r.contentHash = *cs
	if cs.Type.Trailing() {

	rec := httptest.NewRecorder()
	// construct HTTP request for Get Object end point.
	req, err := newTestSignedRequestV4(http.MethodPut, getPutObjectURL("", bucketName, objectName),
		int64(5), bytes.NewReader([]byte("hello")), credentials.AccessKey, credentials.SecretKey, map[string]string{})
	if err != nil {
		t.Fatalf("failed to create HTTP request for Put Object: <ERROR> %v", err)
	}

Das ist auch bei den meisten Webanwendungen der Fall.

Viele, viele Benutzer, aber Ihr Server wartet 🕙 darauf, dass deren nicht so gute Internetverbindungen die <abbr title="Request – Anfrage: Daten, die der Client zum Server sendet">Requests</abbr> übermitteln.

Und dann wieder warten 🕙, bis die <abbr title="Response – Antwort: Daten, die der Server zum anfragenden Client zurücksendet">Responses</abbr> zurückkommen.

func isVeeamSOSAPIObject(object string) bool {
	switch object {
	case systemXMLObject, capacityXMLObject:
		return true
	default:
		return false
	}
}

// isVeeamClient - returns true if the request is from Veeam client.
func isVeeamClient(ctx context.Context) bool {
	ri := logger.GetReqInfo(ctx)
	return ri != nil && strings.Contains(ri.UserAgent, veeamAgentSubstr)
}

// administrator.
//
// - from `Subject.CommonName` field from the STS request for
// AssumeRoleWithCertificate. In this case, the policy for the STS credential
// has the same name as the value of this field.
//
// - from special JWT claim from STS request for AssumeRoleWithOIDC API (when
// not using RoleARN). The claim value can be a string or a list and refers to

pkg net/http, method (*Server) RegisterOnShutdown(func())
pkg net/http, method (*Server) ServeTLS(net.Listener, string, string) error
pkg net/http/fcgi, func ProcessEnv(*http.Request) map[string]string
pkg net/http/httptest, method (*Server) Certificate() *x509.Certificate
pkg net/http/httptest, method (*Server) Client() *http.Client
pkg reflect, func MakeMapWithSize(Type, int) Value

		u.Credentials.AccessKey = user
	}
	if u.Credentials.SessionToken != "" {
		jwtClaims, err := extractJWTClaims(u)
		if err != nil {
			if u.Credentials.IsTemp() {
				// We should delete such that the client can re-request
				// for the expiring credentials.
				deleteKeyEtcd(ctx, ies.client, getUserIdentityPath(user, userType))
				deleteKeyEtcd(ctx, ies.client, getMappedPolicyPath(user, userType, false))
			}
			return nil
		}

///

### <a href="https://requests.readthedocs.io" class="external-link" target="_blank">Requests</a>

**FastAPI** n'est pas réellement une alternative à **Requests**. Leur cadre est très différent.

Il serait en fait plus courant d'utiliser Requests _à l'intérieur_ d'une application FastAPI.

Mais quand même, FastAPI s'est inspiré de Requests.

		Help:      "Number of GET requests proxied to replication target",
		Type:      counterMetric,
	}
}

func getClusterReplProxiedHeadOperationsMD(ns MetricNamespace) MetricDescription {
	return MetricDescription{
		Namespace: ns,
		Subsystem: replicationSubsystem,
		Name:      proxiedHeadRequestsTotal,
		Help:      "Number of HEAD requests proxied to replication target",