{* ../../docs_src/sql_databases/tutorial001_an_py310.py ln[21:22] hl[21:22] *}

### 创建会话（Session）依赖项

**`Session`** 会存储**内存中的对象**并跟踪数据中所需更改的内容，然后它**使用 `engine`** 与数据库进行通信。

我们会使用 `yield` 创建一个 FastAPI **依赖项**，为每个请求提供一个新的 `Session` 。这确保我们每个请求使用一个单独的会话。🤓

然后我们创建一个 `Annotated` 的依赖项 `SessionDep` 来简化其他也会用到此依赖的代码。

{* ../../docs_src/sql_databases/tutorial001_an_py310.py ln[25:30]  hl[25:27,30] *}

{* ../../docs_src/sql_databases/tutorial001_an_py310.py ln[21:22] hl[21:22] *}

### 세션 의존성 생성하기

**`Session`**은 **메모리에 객체**를 저장하고 데이터에 필요한 모든 변경 사항을 추적한 후, **`engine`을 통해** 데이터베이스와 통신합니다.

`yield`를 사용해 FastAPI의 **의존성**을 생성하여 각 요청마다 새로운 `Session`을 제공합니다. 이는 요청당 하나의 세션만 사용되도록 보장합니다. 🤓

그런 다음 이 의존성을 사용하는 코드를 간소화하기 위해 `Annotated` 의존성 `SessionDep`을 생성합니다.

                # Before this was also used for closing dependencies with yield but
                # those now have their own AsyncExitStack, to properly support
                # streaming responses while keeping compatibility with the previous
                # versions (as of writing 0.117.1) that allowed doing
                # except HTTPException inside a dependency with yield.

    If you want this update with the old behavior use [this interceptor][legacy_interceptor].

 *  Fix: Don't crash decompressing web sockets messages. We had a bug where we assumed deflated
    bytes in would always yield deflated bytes out and this isn't always the case!

 *  Fix: Reliably update and invalidate the disk cache on windows. As originally designed our

func getQuorumDisks(disks []StorageAPI, infos []DiskInfo, readQuorum int) (newDisks []StorageAPI) {
	newDisks, _ = getQuorumDiskInfos(disks, infos, readQuorum)
	return newDisks
}

// Will return io.EOF if continuing would not yield more results.
func (er *erasureObjects) listPath(ctx context.Context, o listPathOptions, results chan<- metaCacheEntry) (err error) {
	defer xioutil.SafeClose(results)
	o.debugf(color.Green("listPath:")+" with options: %#v", o)

	VCNT V0.B8, V0.B8                          // 0058200e
	VCNT V0.B16, V0.B16                        // 0058204e
	WFE                                        // 5f2003d5
	WFI                                        // 7f2003d5
	YIELD                                      // 3f2003d5
	//TODO FABD F0, F5, F11                    // abd4a07e
	//TODO VFABD V30.S2, V8.S2, V24.S2         // 18d5be2e
	//TODO VFABS V5.S4, V24.S4                 // b8f8a04e

The closing of a channel is synchronized before a receive that returns a zero value
because the channel is closed.
</p>

<p>
In the previous example, replacing
<code>c &lt;- 0</code> with <code>close(c)</code>
yields a program with the same guaranteed behavior.
</p>

<p class="rule">
A receive from an unbuffered channel is synchronized before the completion of
the corresponding send on that channel.
</p>

<p>

	// is reached and keep reusing the same source string - that
	// should therefore be always resident in the L1 cache - until we
	// have completed the construction of the result.
	// This yields significant speedups (up to +100%) in cases where
	// the result length is large (roughly, over L2 cache size).
	const chunkLimit = 8 * 1024
	chunkMax := n
	if chunkMax > chunkLimit {

   * usually have much noticeable impact. A value of one permits only one thread to modify the cache
   * at a time, but since read operations and cache loading computations can proceed concurrently,
   * this still yields higher concurrency than full synchronization.
   *
   * <p>Defaults to 4. <b>Note:</b>The default may change in the future. If you care about this
   * value, you should always choose it explicitly.
   *

		case toSend = <-c.outQueue:
			if len(toSend) == 0 {
				continue
			}
		}
		if len(queue) < maxMergeMessages && queueSize+len(toSend) < writeBufferSize-1024 {
			if len(c.outQueue) == 0 {
				// Yield to allow more messages to fill.
				runtime.Gosched()
			}
			if len(c.outQueue) > 0 {
				queue = append(queue, toSend)
				queueSize += len(toSend)
				continue
			}
		}