Quick question, I'm using the Visual Studio's testing framework for unit testing. Just wondering what's the difference开发者_C百科 between using the constructor to do initialization work vs. having a method with [TestInitialize()] attribute?
This post gives an overview of the different methods. As you can see, the ctor is called immediately before the ClassInitialize
(only once, of course) and TestInitialize
.
So put stuff that requires code in ClassInitialize
in your TestInitialize
method. Everything that should be set up before ClassInitialize
goes in the ctor.
Obviously, TestInitialize
content will be executed once before each test. The corresponding method to close after each test is TestCleanup
. For classes, use ClassCleanup
. The same thing exists for assemblies as well (AssemblyInitialize/Cleanup
).
Further reading
Conceptually they are they same, as MSTest creates a new instance of your test class before each test execution. However, technically there are a few differences:
- The ctor is called before
TestInitialize
(no surprise as the latter is an instance method). - You have access to
TestContext
inTestInitialize
. - More inheritance scenarios are enabled with
TestInitialize
: https://stackoverflow.com/a/8689398/67824. - You can assign
readonly
fields in the ctor. I think it's pretty important: https://stackoverflow.com/a/45270180/67824.
The ctor
is for initializing the object.
TestInitialize
is for initializing any objects or data needed to run the test.
精彩评论