- Apply that

IMPORTANT: There is an option in **Gradle** for **Automatic Project Synchronization**.
           As convenient as it'd be for the projects to always be perfect this
           tends to add many many seconds to every branch change. Instead, you
           should manually right click on a project and
           **Gradle > Refresh Gradle Project** if the configuration is out of
           date.

                        // Ignore
                    }
                }).start();
            }

            startLatch.countDown();
            assertTrue(endLatch.await(1, TimeUnit.SECONDS));
            assertTrue(response.isReceived());
        }

        @Test
        @DisplayName("Should handle concurrent error notifications")
        void testConcurrentError() throws InterruptedException {

    enableProtocol(protocol)

    server.enqueue(
      MockResponse
        .Builder()
        .trailers(headersOf("t1", "v2"))
        .bodyDelay(1, TimeUnit.SECONDS)
        .body(protocol, "Hello")
        .build(),
    )

    val call = client.newCall(Request(server.url("/")))
    call.execute().use { response ->
      response.close()

* `datetime.time`:
    * 파이썬의 `datetime.time`.
    * 요청과 응답에서 `14:23:55.003`와 같은 ISO 8601 형식의 `str`로 표현됩니다.
* `datetime.timedelta`:
    * 파이썬의 `datetime.timedelta`.
    * 요청과 응답에서 전체 초(seconds)의 `float`로 표현됩니다.
    * Pydantic은 "ISO 8601 time diff encoding"으로 표현하는 것 또한 허용합니다. [더 많은 정보는 문서를 확인하세요](https://docs.pydantic.dev/latest/concepts/serialization/#custom-serializers).
* `frozenset`:
    * 요청과 응답에서 `set`와 동일하게 취급됩니다:

			if err != nil {
				log.Fatalf("Error reading session policy file: %v", err)
			}
			policy = string(bs)
		}
		stsOpts.Policy = policy
	}
	if expiryDuration != 0 {
		stsOpts.DurationSeconds = int(expiryDuration.Seconds())
	}
	li, err := cr.NewSTSAssumeRole(stsEndpoint, stsOpts)
	if err != nil {
		log.Fatalf("Error initializing STS Identity: %v", err)
	}

	stsEndpointURL, err := url.Parse(stsEndpoint)
	if err != nil {

                    }
                } finally {
                    latch.countDown();
                }
            });
        }

        // Then
        assertTrue(latch.await(10, TimeUnit.SECONDS), "All threads should complete");
        executor.shutdown();
        assertEquals(0, collisions.get(), "There should be no nonce collisions");

import static com.google.common.util.concurrent.Futures.immediateFuture;
import static java.util.Arrays.asList;
import static java.util.concurrent.TimeUnit.MILLISECONDS;
import static java.util.concurrent.TimeUnit.SECONDS;
import static org.junit.Assert.assertThrows;

import com.google.common.cache.CacheLoader.InvalidCacheLoadException;
import com.google.common.cache.TestingCacheLoaders.CountingLoader;

使用 `docker compose` 時這會特別明顯。技術細節請見這段 Docker Compose 常見問題：[為什麼我的服務要花 10 秒才重新建立或停止？](https://docs.docker.com/compose/faq/#why-do-my-services-take-10-seconds-to-recreate-or-stop)

#### 目錄結構 { #directory-structure }

你現在應該會有如下的目錄結構：

```
.
├── app
│   ├── __init__.py
│   └── main.py
├── Dockerfile
└── requirements.txt
```

            try {
              proceed.await(5, TimeUnit.SECONDS)
            } catch (e: InterruptedException) {
              throw RuntimeException(e)
            }
            chain.proceed(chain.request())
          },
        ).build()
    val t1 = makeSynchronousCall(client.newCall(newRequest("http://a/3")))
    ready.await(5, TimeUnit.SECONDS)
    executor.finishJob("http://a/2")

        }
      },
    )

    // This should fail the Call, but not cause an unhandled Exception bubbling up
    client.dispatcher.executorService.shutdownNow()

    val exception = callFailure.get(5, TimeUnit.SECONDS)
    assertThat(exception.message)
      .isNotNull()
      .startsWith("canceled due to")
    assertThat(exception).isInstanceOf<IOException>()
    assertThat(exception.cause)
      .isNotNull()