PHPUnit测试问题-如何对我的班级进行单元测试-PHPUnitTestQuestion-HowtoUnittestmyclass

9  

Unit tests should be written to evaluate the public interface of a class. Your test case should use the class as you intend it to be used in your program. The idea here is to test the behavior (either expected, unexpected, or edge conditions) of the class.

应编写单元测试来评估类的公共接口。您的测试用例应该使用该类,因为您打算在程序中使用它。这里的想法是测试类的行为(预期,意外或边缘条件)。

Both ideas you posted are correct. In theory, you should have enough test cases (routes through your code) that all your methods in the class are run.

您发布的两个想法都是正确的。理论上,您应该有足够的测试用例(通过代码的路由),以便运行类中的所有方法。

As was mentioned, 100% test coverage is a nice goal, but not always realistic.

如前所述,100%的测试覆盖率是一个很好的目标,但并不总是现实的。

Also, in the case of GD, be careful about writing unit tests that test GD's functionality (it's already been tested, you don't need to waste time testing it again). I would read up on using PHPUnit's mocks and stubs (and mocking the filesystem) in the PHPUnit manual.

此外,在GD的情况下,要小心编写测试GD功能的单元测试(它已经过测试,你不需要浪费时间再次测试它)。我会在PHPUnit手册中阅读使用PHPUnit的模拟和存根(以及模拟文件系统)。

Here's what an example test might look like:

以下是示例测试的样子:

public function testImageIsResized()
{
    $image = new Image();
    $image->loadFromPath('some/path');
    $image->resize(200, 300);
    $this->assertEquals(200, $image->getWidth());
    $this->assertEquals(300, $image->getHeight());
}

Now, depending on the expected behavior of the image class, this test might pass without issue, or it might fail because it was expecting the new dimensions to be proportionally constrained to the original image dimensions. But we did not explicitly call the internal method that checks for that constraint in the test itself.

现在,根据图像类的预期行为,这个测试可能通过没有问题,或者因为它期待新的尺寸成比例地限制原始图像尺寸也可能会失败。但是我们没有显式调用在测试本身中检查该约束的内部方法。

5  

You can use the covers annotation to specify if a test covers multiple methods. So if one of your methods calls another method, you can simply add the annotation to the test's docblock and it will be added to your code coverage statistics, e.g.

您可以使用封面注释来指定测试是否涵盖多种方法。因此,如果您的某个方法调用另一个方法,您只需将注释添加到测试的docblock中,它就会被添加到您的代码覆盖率统计信息中,例如:

/**
 * @test
 * @covers MyClass::something()
 * @covers MyClass::_somethingElse()
 */
public function somethingWorksAsExpected()
{
    $this->assertSame($expected, $this->testObject->something());
}

For personal projects, 100% Code coverage is fine. However, I've seen talks at conferences where 100% are doubted to be necessary. Despite all the benefits, tests take time to write and in a budgeted project it might be sufficient to just test 80/20 and leave out uncritical low priority features of your app.

对于个人项目,100%的代码覆盖率是好的。但是,我在会议上看到过100%被怀疑是必要的会议。尽管有各种好处,但测试需要时间来编写,在预算项目中,仅测试80/20并省略应用程序的非关键性低优先级功能就足够了。

As for how to test your class, have a look at the chapter on Behaviour Driven Development in the PHPUnit Manual. In your case, I'd test the functionality you described in your question.

至于如何测试你的类,请参阅PHPUnit手册中关于行为驱动开发的章节。在您的情况下,我会测试您在问题中描述的功能。

0  

Stephen Melrose said:

斯蒂芬梅尔罗斯说:

However, some of the methods run others (rightly so may I add), so you then have a chain of dependency. But I also read that each Unit test should be independent from the other

但是,有些方法会运行其他方法(正确地说,我可以添加),因此您将拥有一系列依赖关系。但我也读到每个单元测试应该独立于另一个

Test independence isn't about not testing the same code twice, it's about whether the result (or lack of result) of one test affects the result of another. If your first test inserts some data and then fails before it can remove that data, your second test might get different results than expected. Ideally you should be able to run your tests in a random order, or run some and not others.

测试独立性不是关于不测试相同代码两次,而是关于一个测试的结果(或缺少结果)是否会影响另一个测试的结果。如果您的第一个测试插入一些数据然后在它可以删除该数据之前失败,那么您的第二个测试可能会得到与预期不同的结果。理想情况下,您应该能够以随机顺序运行测试,或运行一些而不是其他运行。