Published on 17 June 2013.
Today’s though is about how to know when test coverage is good enough.
One benefit of having good test coverage is that you are less afraid to modify code. It is powerful to be able to check the correctness of your implementation by running a suit of tests with a single press of a button. But if you are going to trust your tests that much, it is important that the coverage is relatively good. But what is good enough?
I suggest two activities that can help you answer that question:
In the first activity you familiarize yourself with the existing tests (and also the implementation) and learn what is there. You might think of cases that are not there that you can add. You might not get a good overview of what are actually tested, so you refactor the tests to more clearly reflect that.
In the second activity you might be able to break the implementation in a way that your tests do not catch. In that case, you have found a test case that you might want to write. Otherwise, the test coverage is probably good enough.
I think you need different coverage in different areas of the code to feel confident when making a change. Instead of saying that code coverage should be a specific percentage, you can do these two activities to help you convince yourself that coverage is good enough.
Site proudly generated by Hakyll.