问题描述
在PHP中魔术方法的单元测试实现时，推荐使用哪些方法来调用这些方法?
When it comes to unit-testing implementations of magic methods in PHP, what is the recommended means of invoking those methods?
我看到三个可用选项:
-
显式/直接调用它们:
Invoking them explicitly/directly:
$object->__get('someValue');
间接调用它们(使用旨在触发它们的任何操作):
Invoking them indirectly (using whatever action is intended to trigger them):
$object->someValue; \\ Where __get() is implemented.
使用这两种方法调用它们.
Invoking them using both methods.
是否有任何单元测试的资深人士可以解释哪个(如果有的话)将是显而易见的选择，为什么会这样?
Are there any unit testing veterans that could explain which (if any) would be the obvious choice, and why that might be?
(这可能是在主观/辩论领域附近跳舞的，但是我希望问一下在解决这个问题时我应该考虑一些公认的原则.)
推荐答案
您应该测试可观察的行为.因此，第二个($obj->property) 必须经过测试，没问题.
You should be testing observable behavior. So, the second one ($obj->property) must be tested, no question.
至于直接直接调用getter，更多的是判断.一旦您说了__get('someProperty')，在我看来，您几乎已经陷入僵局，这是一个神奇的属性.由于我一直尝试不更改测试，因此如果出于某种原因我希望它成为普通的非魔术属性，会使事情复杂化很多.从好的方面来说，您可以调用getter并以另一种方式获取该属性，并确保它们都具有完全相同的结果.
As for actually calling the getter directly, that's more of a judgement call. Once you have said __get('someProperty'), in my opinion you've pretty much set in stone that it is a magic property. Since I try to never change the tests, that complicates things quite a bit if, for some reason, I want that to be a plain old non-magic property. On the plus side, you can call the getter and get the property the other way as well and ensure they both have the exact same result.
这篇关于单元测试魔术方法的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！