// transition to DONE
  public void testInterruptThrows() throws Exception {
    CountDownLatch isInterruptibleRegistered = new CountDownLatch(1);
    SettableFuture<String> taskResult = SettableFuture.create();
    InterruptibleTask<String> task =
        new InterruptibleTask<String>() {
          @Override
          String runInterruptibly() throws Exception {
            BrokenChannel bc = new BrokenChannel();

* 它可以对该**响应**做一些事情或者执行任何需要的代码。
* 然后它返回这个**响应**。

/// note | 技术细节

如果你有使用 `yield` 的依赖，依赖中的退出代码会在中间件之后运行。

如果有任何后台任务（会在[后台任务](background-tasks.md)一节中介绍，你稍后会看到），它们会在所有中间件之后运行。

///

## 创建中间件 { #create-a-middleware }

要创建中间件，你可以在函数的顶部使用装饰器 `@app.middleware("http")`。

中间件函数会接收：

* `request`。
* 一个函数 `call_next`，它会把 `request` 作为参数接收。
    * 这个函数会把 `request` 传递给相应的*路径操作*。
    * 然后它返回由相应*路径操作*生成的 `response`。

     * <p>
     * Determining the transitive deps of a dependency which has been resolved as
     * non-transitive is difficult because the process of resolving removes the
     * transitive deps. To sidestep this issue, we create a configuration per
     * direct dependency version. This specially named and unique configuration
     * will contain all of the transitive dependencies of this particular

///

### 2 단계: `FastAPI` "인스턴스" 생성 { #step-2-create-a-fastapi-instance }

{* ../../docs_src/first_steps/tutorial001_py310.py hl[3] *}

여기에서 `app` 변수는 `FastAPI` 클래스의 "인스턴스"가 됩니다.

이것은 여러분의 모든 API를 생성하기 위한 상호작용의 주요 지점이 될 것입니다.

### 3 단계: *경로 처리* 생성 { #step-3-create-a-path-operation }

#### 경로 { #path }

여기서 "경로"는 첫 번째 `/`부터 시작하는 URL의 뒷부분을 의미합니다.

        // Web
        for (final WebConfig webConfig : webConfigList) {
            final String sid = ComponentUtil.getCrawlingConfigHelper().store(sessionId, webConfig);

            // create crawler
            final Crawler crawler = ComponentUtil.getComponent(Crawler.class);
            crawler.setSessionId(sid);
            sessionIdList.add(sid);

            final String urlsStr = webConfig.getUrls();

   * ordering of all views in the returned table. Note that some views of the original table and the
   * copied table may have different iteration orders. For more control over the ordering, create a
   * {@link Builder} and call {@link Builder#orderRowsBy}, {@link Builder#orderColumnsBy}, and
   * {@link Builder#putAll}
   *

        return new SimpleBuildEventListener(writer);
    }

    protected final void createTerminal(C context) {
        if (context.terminal == null) {
            // Create the build log appender; also sets MavenSimpleLogger sink
            ProjectBuildLogAppender projectBuildLogAppender =
                    new ProjectBuildLogAppender(determineBuildEventListener(context));

    @Execute
    public JsonResponse<ApiResult> post$setting(final String dictId, final CreateBody body) {
        body.dictId = dictId;
        validateApi(body, messages -> {});
        body.crudMode = CrudMode.CREATE;
        final StemmerOverrideItem entity = createStemmerOverrideItem(this, body, () -> {
            throwValidationErrorApi(messages -> messages.addErrorsCrudFailedToCreateInstance(GLOBAL));
            return null;

   * [computeIfAbsent] implementation. No locks are held while calling this function.
   */
  fun <T : Any> tag(
    type: Class<T>,
    computeIfAbsent: () -> T,
  ): T

  /**
   * Create a new, identical call to this one which can be enqueued or executed even if this call
   * has already been.
   *
   * The tags on the returned call will equal the tags as on [request]. Any tags that were computed

   * loaded by this test and that will obviously remain loaded for as long as the test is running.
   * So in order to check ClassLoader garbage collection we need to create a new ClassLoader and
   * make it load its own version of FinalizableReferenceQueue. Then we need to interact with that
   * parallel version through reflection in order to exercise the parallel