问题

测试场景是从一个名为异步调用的模拟客户端服务中获取结果。在这里，如果我尝试单独运行上述测试，它们会成功执行。但是当尝试通过```mvn clean install```构建项目时，成功的测试用例会出错。

如何处理异常情况，以使其不影响预期的方法行为测试用例？或者可能是什么原因导致多个异步调用出现问题？
如何处理/关闭生成的线程，或者在开始下一个客户端调用之前等待其完成？

// 服务类
public String serviceMethod(String arg1, String arg2) {
return mockedclient.fetchDetails(arg1, arg2)
.toCompletableFuture()
.thenapplyasync(HashSet::new) // 假设收到的响应是一个列表
.get(3,TimeUnit.SECONDS);
}

// 测试类
def "应该返回从客户端检索到的结果" () {
given:
def arg1 = "value1"
def arg2 = "value2"

def futureList = CompletableFuture.completedFuture(Collections.toList("RESULT"))
def futureSet = new BlockingVariable<Set>()

when:
def result = serviceMethod(arg1, arg2)

then:
1 * mockedClient.fetchDetails() >> futureList
0 * futureList.thenApplyAsync() >> futureSet
0 * futureSet.get(,) >> Collections.toSet("RESULT")
}

def "应该在从客户端检索结果时返回空响应" () {
given:
def arg1 = "value1"
def arg2 = "value2"

def futureList = CompletableFuture.completedFuture(Collections.toList("RESULT"))
def futureSet = new BlockingVariable<Set>()

when:
def result = serviceMethod(arg1, arg2)

then:
1 * mockedClient.fetchDetails() >> CompletableFuture.runAsync( { throw new Exception("无法获取经纪人分析师详情") } )
0 * futureList.thenApplyAsync() >> futureSet
0 * futureSet.get(,) >> Collections.toSet("RESULT")
}

<details>
<summary>英文:</summary>

The test scenario is to get the result from a mocked client service called asynchronously. Here, if I try and run the mentioned tests individually, they execute successfully. But when trying to build the project via ```mvn clean install``` , the success test case gives error. 

 How do I handle the exception case scenario so that it does not affect the expected method behavior test cases? Or possibly what is causing the issue with multiple async calls? 
How do I handle/shutdown the spawned up thread, or wait for its completion before starting the next client call?

// SERVICE CLASS
public String serviceMethod(String arg1, String arg2) {
return mockedclient.fetchDetails(arg1, arg2)
.toCompletableFuture()
.thenapplyasync(HashSet::new) // assuming the received response is a list
.get(3,TimeUnit.SECONDS);
}

//TEST CLASS
def " should return result retrieved from the client" () {
given:
def arg1 = "value1"
def arg2 = "value2"

def futureList = CompletableFuture.completedFuture(Collections.toList("RESULT"))
def futureSet = new BlockingVariable<Set<String>>()

when:
def result = serviceMethod(arg1, arg2)

then:
1 * mockedClient.fetchDetails() >> futureList
0 * futureList.thenApplyAsync() >> futureSet
0 * futureSet.get(,) >> Collections.toSet("RESULT")
}

def " should return empty response when result retrieved from the client" () {
given:
def arg1 = "value1"
def arg2 = "value2"

def futureList = CompletableFuture.completedFuture(Collections.toList("RESULT"))
def futureSet = new BlockingVariable<Set<String>>()

when:
def result = serviceMethod(arg1, arg2)

then:
1 * mockedClient.fetchDetails() >> CompletableFuture.runAsync( { throw new Exception("Failed to fetch broker Analyst details " } )
0 * futureList.thenApplyAsync() >> futureSet
0 * futureSet.get(,) >> Collections.toSet("RESULT")
}

</details>


# 答案1
**得分**: 0

这不是一个完整的回答，因为您也没有提供完整的[MCVE](https://stackoverflow.com/help/mcve)。因此，我无法通过复制、编译和运行您的代码片段来重现您的问题。但有一些事情让我感到奇怪：

* 您只是说某些东西不起作用，但没有显示任何错误消息和堆栈跟踪。

* 您的生产服务类使用了`mockedclient`，至少名称很奇怪。为什么在生产中使用模拟，或者使用模拟来测试某些东西呢？

* `futureList`在两个特性方法中都不是模拟，而是一个真实的对象。因此，您不能对其进行交互检查和返回类似于`0 * futureList.thenApplyAsync(_) >> futureSet`的存根结果。

* `futureSet`在两个特性方法中也不是模拟，而是一个真实的对象。因此，您不能对其进行交互检查和返回类似于`0 * futureSet.get(_, _) >> Collections.toSet("RESULT")`的存根结果。

* 在两个特性方法中，您使用了一个名为`mockedClient`的变量，但忘记了显示它在哪里以及如何定义/初始化。

我猜您的测试不仅在Maven上抛出错误，而且在本地也是如此，因为它们包含错误。

<details>
<summary>英文:</summary>

This is not a full answer because you didn&#39;t provide a full [MCVE](https://stackoverflow.com/help/mcve) either. So I cannot reproduce your problem by copying, compiling and running your code fragments. But a few things strike me as odd:

* You just say something is not working but do not show any error messages and stacktraces.

* Your production service class uses `mockedclient`, at least the name is strange. Why would you use a mock in production or test something using a mock?

* `futureList` in both feature methods is not a mock but a real object. Thus, you cannot check interactions and return stubbed results like `0 * futureList.thenApplyAsync(_) &gt;&gt; futureSet` on it.

* `futureSet` in both feature methods is not a mock but a real object. Thus, you cannot check interactions and return stubbed results like `0 * futureSet.get(_,_) &gt;&gt; Collections.toSet(&quot;RESULT&quot;)` on it.

* In both feature methods you use a variable `mockedClient` but forgot to show where and how it is defined/initialised.

I assume your tests are throwing errors not just on Maven but also locally because, well - they contain bugs.

</details>