# Dependências com yield { #dependencies-with-yield }

O **FastAPI** possui suporte para dependências que realizam <dfn title='às vezes também chamado de "código de saída", "código de limpeza", "código de teardown", "código de fechamento", "código de saída do gerenciador de contexto", etc.'>alguns passos extras ao finalizar</dfn>.

Para fazer isso, utilize `yield` em vez de `return`, e escreva os passos extras (código) depois.

/// tip | Dica

# Abhängigkeiten mit `yield` { #dependencies-with-yield }

FastAPI unterstützt Abhängigkeiten, die einige <dfn title="manchmal auch genannt: „Exit Code“, „Cleanup Code“, „Teardown Code“, „Closing Code“, „Kontextmanager Exit Code“, usw.">zusätzliche Schritte nach Abschluss</dfn> ausführen.

Verwenden Sie dazu `yield` statt `return` und schreiben Sie die zusätzlichen Schritte / den zusätzlichen Code danach.

/// tip | Tipp

* Bump container-vm version in config-test.sh ([#35705](https://github.com/kubernetes/kubernetes/pull/35705), [@mtaufen](https://github.com/mtaufen))
* Delete all firewall rules (and optionally network) on GCE/GKE cluster teardown ([#34577](https://github.com/kubernetes/kubernetes/pull/34577), [@ixdy](https://github.com/ixdy))

* Make NodeController recognize deletion tombstones ([#34786](https://github.com/kubernetes/kubernetes/pull/34786), [@davidopp](https://github.com/davidopp))
* Delete all firewall rules (and optionally network) on GCE/GKE cluster teardown ([#34577](https://github.com/kubernetes/kubernetes/pull/34577), [@ixdy](https://github.com/ixdy))

  private var client = clientTestRule.newClient()
  private var cache: Cache? = null

  @BeforeEach
  fun setUp() {
    server.protocolNegotiationEnabled = false
  }

  @AfterEach
  fun tearDown() {
    java.net.Authenticator.setDefault(null)
    System.clearProperty("proxyHost")
    System.clearProperty("proxyPort")
    System.clearProperty("http.proxyHost")
    System.clearProperty("http.proxyPort")

","trigger","isImmediatePropagationStopped","handleObjIn","eventHandle","events","t","handleObj","handlers","namespaces","origType","elemData","create","handle","triggered","dispatch","bindType","delegateCount","setup","mappedTypes","origCount","teardown","removeEvent","nativeEvent","handlerQueue","fix","delegateTarget","preDispatch","isPropagationStopped","currentTarget","rnamespace","postDispatch","matchedHandlers","matchedSelectors","addProp","hook","Event","enumerable","originalEvent","writa...

                # Enter the SSE context manager on the request-scoped
                # exit stack. The stack outlives the streaming response,
                # so __aexit__ runs via proper structured teardown, not
                # via GeneratorExit thrown into an async generator.
                sse_receive_stream = await async_exit_stack.enter_async_context(
                    _sse_producer_cm()
                )

      directory = "/cache".toPath(),
      maxSize = Int.MAX_VALUE.toLong(),
    )

  @BeforeEach
  fun setUp() {
    platform.assumeNotOpenJSSE()
  }

  @AfterEach
  fun tearDown() {
    cache.close()
    fileSystem.checkNoOpenFiles()
  }

  @Test
  fun get() {
    server.enqueue(
      MockResponse
        .Builder()
        .body("abc")
        .clearHeaders()

","trigger","isImmediatePropagationStopped","handleObjIn","eventHandle","events","t","handleObj","handlers","namespaces","origType","elemData","create","handle","triggered","dispatch","bindType","delegateCount","setup","mappedTypes","origCount","teardown","removeEvent","nativeEvent","handlerQueue","fix","delegateTarget","preDispatch","isPropagationStopped","currentTarget","rnamespace","postDispatch","matchedHandlers","matchedSelectors","addProp","hook","Event","enumerable","originalEvent","writa...

- Node shutdown controller made a best effort to wait for CSI Drivers to complete the volume teardown process according to the pod priority groups. ([#125070](https://github.com/kubernetes/kubernetes/pull/125070), [@torredil](https://github.com/torredil)) [SIG Node, Storage and Testing]