그리고 `Depends(scope="request")`(기본값)를 사용하면, `yield` 이후의 종료 코드는 응답이 전송된 후에 실행됩니다.

자세한 내용은 [`yield`가 있는 의존성 - 조기 종료와 `scope`](../tutorial/dependencies/dependencies-with-yield.md#early-exit-and-scope) 문서를 참고하세요.

### `yield`가 있는 의존성과 `StreamingResponse`, 기술 세부사항 { #dependencies-with-yield-and-streamingresponse-technical-details }

  }

  private fun logWithTime(message: String) {
    val startNs = startNs
    val timeMs =
      if (startNs == null) {
        // Event occurred before start, for an example an early cancel.
        0L
      } else {
        TimeUnit.NANOSECONDS.toMillis(System.nanoTime() - startNs)
      }

    logger.invoke("[$timeMs ms] $message")
  }

而当使用默认的 `Depends(scope="request")` 时，`yield` 之后的退出代码会在响应发送之后执行。

你可以在文档 [带 `yield` 的依赖项 - 提前退出与 `scope`](../tutorial/dependencies/dependencies-with-yield.md#early-exit-and-scope) 中了解更多。

### 带 `yield` 的依赖项与 `StreamingResponse`（技术细节） { #dependencies-with-yield-and-streamingresponse-technical-details }

在 FastAPI 0.118.0 之前，如果你使用带 `yield` 的依赖项，它会在*路径操作函数*返回后、发送响应之前运行 `yield` 之后的退出代码。

Puedes leer más al respecto en la documentación de [Dependencias con `yield` - Salida temprana y `scope`](../tutorial/dependencies/dependencies-with-yield.md#early-exit-and-scope).

### Dependencias con `yield` y `StreamingResponse`, detalles técnicos { #dependencies-with-yield-and-streamingresponse-technical-details }

==============================================================================*/
#ifndef TENSORFLOW_C_EAGER_IMMEDIATE_EXECUTION_TENSOR_HANDLE_H_
#define TENSORFLOW_C_EAGER_IMMEDIATE_EXECUTION_TENSOR_HANDLE_H_

#include "tensorflow/c/eager/abstract_tensor_handle.h"
#include "tensorflow/c/tensor_interface.h"
#include "tensorflow/core/framework/types.pb.h"
#include "tensorflow/core/platform/status.h"

namespace tensorflow {

///

## Early exit and `scope` { #early-exit-and-scope }

Normally the exit code of dependencies with `yield` is executed **after the response** is sent to the client.

// algorithms compute correct checksums. If any algorithm
// produces an incorrect checksum it fails with a hard error.
//
// bitrotSelfTest tries to catch any issue in the bitrot implementation
// early instead of silently corrupting data.
func bitrotSelfTest() {
	checksums := map[BitrotAlgorithm]string{
		SHA256:          "a7677ff19e0182e4d52e3a3db727804abc82a5818749336369552e54b838b004",

#ifndef TENSORFLOW_C_EAGER_ABSTRACT_OPERATION_H_
#define TENSORFLOW_C_EAGER_ABSTRACT_OPERATION_H_

#include <memory>

#include "absl/types/span.h"
#include "tensorflow/c/eager/abstract_tensor_handle.h"
#include "tensorflow/c/tensor_interface.h"
#include "tensorflow/core/framework/tensor_shape.h"
#include "tensorflow/core/framework/types.pb.h"
#include "tensorflow/core/platform/status.h"

import org.apache.maven.api.annotations.Nonnull;
import org.apache.maven.api.annotations.Nullable;

/**
 * Defines a simple logging interface for Maven CLI operations. These operations happen "early", when there may
 * be no logging set up even. Implementations may be "accumulating", in which case {@link #drain()}  method should
 * be used.
 * <p>

## Tips for Success

- **Small PRs**: Easier to review, faster to merge. Split large changes logically.
- **Clear Commits**: Use `git rebase -i` to refine history before submitting.
- **Engage Early**: Discuss complex changes in issues or Slack (https://slack.min.io) before coding.
- **Be Responsive**: Address reviewer feedback promptly to keep PRs moving.
- **Learn from Reviews**: Use feedback to improve future contributions.