## Fixe a sua versão de `fastapi` { #pin-your-fastapi-version }

A primeira coisa que você deve fazer é "fixar" a versão do **FastAPI** que você está utilizando na versão mais recente específica que você sabe que funciona corretamente para a sua aplicação.

  /**
   * Returns true if the host and port are unchanged from when this was created. This is used to
   * detect if followups need to do a full connection-finding process including DNS resolution, and
   * certificate pin checks.
   */
  fun sameHostAndPort(url: HttpUrl): Boolean

  /**
   * A plan holds either an immediately-usable connection, or one that must be connected first.

      errorCode: ErrorCode,
    ) {
      check(type == -1)
      this.type = Http2.TYPE_RST_STREAM
      this.streamId = streamId
      this.errorCode = errorCode
    }

    override fun ping(
      ack: Boolean,
      payload1: Int,
      payload2: Int,
    ) {
      check(type == -1)
      type = Http2.TYPE_PING
      this.ack = ack
      this.payload1 = payload1

经常添加新功能，定期修复错误，并且代码仍在持续改进。

这就是为什么当前版本仍然是`0.x.x`，这反映出每个版本都可能有Breaking changes。 这遵循[语义版本控制](https://semver.org/)的约定。

你现在就可以使用 **FastAPI** 创建生产环境应用程序（你可能已经这样做了一段时间），你只需确保使用的版本可以与其余代码正确配合即可。

## 固定你的 `fastapi` 版本 { #pin-your-fastapi-version }

你应该做的第一件事是将你正在使用的 **FastAPI** 版本“固定”到你知道适用于你的应用程序的特定最新版本。

例如，假设你在应用程序中使用版本`0.112.0`。

如果你使用`requirements.txt`文件，你可以使用以下命令指定版本：

```txt
fastapi[standard]==0.112.0
```

	if m.outBlock == nil {
		// Closed
		return
	}
	select {
	case m.outBlock <- struct{}{}:
	default:
		gridLogIf(m.ctx, errors.New("output unblocked overflow"))
	}
}

func (m *muxServer) ping(seq uint32) pongMsg {
	if !m.checkSeq(seq) {
		msg := fmt.Sprintf("receive sequence number mismatch. want %d, got %d", m.RecvSeq, seq)
		return pongMsg{Err: &msg}
	}
	select {
	case <-m.ctx.Done():

        SmbPipeInputStream in1 = handle.getInput();
        SmbPipeInputStream in2 = handle.getInput();
        SmbPipeOutputStream out1 = handle.getOutput();
        SmbPipeOutputStream out2 = handle.getOutput();

        // Assert: same instances are returned subsequently
        assertSame(in1, in2, "Input stream should be cached");

{* ../../docs_src/server_sent_events/tutorial005_py310.py hl[14] *}

## 技術詳細 { #technical-details }

FastAPI は SSE のいくつかのベストプラクティスを標準で実装しています。

- メッセージがない場合は 15 秒ごとに「キープアライブ」用の `ping` コメントを送信し、一部のプロキシが接続を閉じるのを防ぎます（[HTML 仕様: Server-Sent Events](https://html.spec.whatwg.org/multipage/server-sent-events.html#authoring-notes) の推奨に従います）。
- ストリームの**キャッシュを防止**するため、`Cache-Control: no-cache` ヘッダーを設定します。

#### Do-stuff-later pool

Sometimes there's an action required like calling the application layer or responding to a ping, and the thread discovering the action is not the thread that should do the work. We enqueue a runnable on this executor and it gets handled by one of the executor's threads.

### Locks

    assertThat(sendPingFrame(true, expectedPayload1, expectedPayload2)).isEqualTo(frame)
    reader.nextFrame(
      requireSettings = false,
      object : BaseTestHandler() {
        override fun ping(
          ack: Boolean,
          payload1: Int,
          payload2: Int,
        ) {
          assertThat(ack).isTrue()
          assertThat(payload1).isEqualTo(expectedPayload1)

      if (delegate == null) {
        return;
      }

      /*
       * If we're about to complete the TimeoutFuture, we want to release our reference to it.
       * Otherwise, we'll pin it (and its result) in memory until the timeout task is GCed. (The
       * need to clear our reference to the TimeoutFuture is the reason we use a *static* nested
       * class with a manual reference back to the "containing" class.)