This is my current understanding of the problem. Please provide feedback whether I'm missing something.

We have lots of top-level functions and it's often unclear

• issue1: whether they belong to the functional core (not the imperative shell) and therefore needs to be tested
• issue2: whether they are currently being tested
• issue3: if not, what kind of simple tests we can add as a starting point

For example with the function getArtifactNameToConstantMapping()

• issue1: definitely part of the functional core and we relay on it for multiple features
• issue2: it's not being tested currently. It's used in other tests which is not the same thing.
• issue3: the function works so we should not modify it but simply write what Martin Fowler calls acceptance tests for it. Meaning: if we give her this kind of objects as input, this is the output we currently get.

Regarding issue 2, an issue is that if you have a class FindKitten, there is a strong convention that you test it in the class FikkndKittenTest. And you can jump there with the action Go To Test.

But there is no such convention for top-level functions AFAIK.