如何在MS Test结合Shouldly的单元测试中断言邮件发送

我有一个发送邮件的SendEmail方法，代码如下：

public async Task SendEmail(string Id, string exception)
{
    SendEmailModel sendEmailmodel = new SendEmailModel();
    var emailAddress = _emailConfiguration.ToEmailAddress;
    string[] toAddresses = emailAddress.Split(',', 10, StringSplitOptions.RemoveEmptyEntries);
    sendEmailmodel.Subject = "Subject";
    sendEmailmodel.ToAddresses = toAddresses;
    sendEmailmodel.Message = exception;
    await _emailService.SendEmailAsync(sendEmailmodel);
}

我正在使用MS Test结合Shouldly编写单元测试验证该方法，尝试写了如下测试方法但不知如何添加断言：

[TestMethod]
public async Task SendEmail_should_send_email()
{
    SendEmailModel sendEmailmodel = new SendEmailModel();
    var emailAddress = "kumara@gmail.com";
    string[] toAddresses = emailAddress.Split(',', 10, StringSplitOptions.RemoveEmptyEntries);
    sendEmailmodel.Subject = "HA test email";
    sendEmailmodel.ToAddresses = toAddresses;
    sendEmailmodel.Message = "test";
    await emailService.SendEmailAsync(sendEmailmodel);

    A.CallTo(() => this.emailService.SendEmailAsync(sendEmailmodel));

   //how to assert 
   await this.limitsLogic.SendEmail("120", "test");
    
}

解决方案 #

单元测试的核心是验证依赖服务（这里是_emailService）是否被正确调用，且传递的参数符合预期。结合你使用的FakeItEasy（从A.CallTo语法判断）和Shouldly，按以下步骤调整测试：

1. 正确初始化测试依赖

• 模拟_emailConfiguration，设置测试用的目标邮箱地址
• 模拟IEmailService接口实例，注入到被测的limitsLogic类中

2. 修正测试流程

• 先执行被测方法limitsLogic.SendEmail
• 再断言依赖服务的调用行为和参数正确性

方式一：用FakeItEasy直接断言参数匹配

[TestMethod]
public async Task SendEmail_should_send_email_with_correct_parameters()
{
    // 准备测试依赖
    var testConfig = new EmailConfiguration
    {
        ToEmailAddress = "kumara@gmail.com"
    };
    var fakeEmailService = A.Fake<IEmailService>();
    var limitsLogic = new LimitsLogic(testConfig, fakeEmailService); // 替换为你的实际类构造

    // 执行被测方法
    await limitsLogic.SendEmail("120", "test");

    // 断言邮箱服务被调用，且参数完全符合预期
    A.CallTo(() => fakeEmailService.SendEmailAsync(A<SendEmailModel>.That.Matches(model =>
        model.Subject == "Subject" &&
        model.ToAddresses.SequenceEqual(new[] { "kumara@gmail.com" }) &&
        model.Message == "test"
    ))).MustHaveHappenedOnceExactly();
}

方式二：捕获参数后用Shouldly做细致断言

这种方式更贴合Shouldly的语法风格，能精准验证每个参数字段：

[TestMethod]
public async Task SendEmail_should_send_email_with_correct_parameters()
{
    // 准备测试依赖
    var testConfig = new EmailConfiguration
    {
        ToEmailAddress = "kumara@gmail.com"
    };
    var fakeEmailService = A.Fake<IEmailService>();
    var limitsLogic = new LimitsLogic(testConfig, fakeEmailService);

    // 捕获调用时的参数
    SendEmailModel sentModel = null;
    A.CallTo(() => fakeEmailService.SendEmailAsync(A<SendEmailModel>._))
        .Invokes((SendEmailModel model) => sentModel = model);

    // 执行被测方法
    await limitsLogic.SendEmail("120", "test");

    // 用Shouldly断言参数细节
    sentModel.ShouldNotBeNull();
    sentModel.Subject.ShouldBe("Subject");
    sentModel.ToAddresses.ShouldBe(new[] { "kumara@gmail.com" });
    sentModel.Message.ShouldBe("test");
}

关键注意点 #

• 不要在测试中主动调用emailService.SendEmailAsync，这是被测方法的职责
• 测试的核心是验证被测方法是否按预期与依赖交互，而非验证邮件是否真的发送成功（那是集成测试的范畴）
• 确保模拟的依赖（配置、邮箱服务）正确注入到被测类中，避免空引用或真实服务调用

内容的提问来源于stack exchange，提问作者Kumara