Write tests to improve code coverage

[danielabrozzoni]

Our CI runs on master a special job called "Code coverage". The goal of this job is to calculate which lines of code are covered by our tests, and which ones aren't, and publish the result as a code coverage report. You have two ways of viewing the report:

1. On the coveralls website. Just go on https://coveralls.io/github/bitcoindevkit/bdk?branch=master

2. Download from the CI artifacts:

Go to the latest commit, click on the "Code Coverage / Code Coverage" job:

You then click on the job summary:

And you click to download the coverage-report:

---

Scroll around - you'll eventually encounter a big chunk of red code, such as:

Your goal is now to write a test that hits that piece of code and opening a PR. Easy* :)

• Easy, but extremely important! The higher code coverage we have, the better tests we have, the higher the chance that we catch bugs before they slip in. Your contribution here is extremely valuable!

[vladimirfomene]

@danielabrozzoni, I'm taking up this one.

[danielabrozzoni]

Thanks @vladimirfomene! I'm not specifically assigning anyone to this issue as it's quite open ended: one person can't really improve the code coverage from our current 79% to >90% by themselves, can it? 😄
Feel free to open as many PRs as you want, adding as many tests as you want, and move to something else when you feel like so :)

[LLFourn]

I think the code coverage tool should be used to find non-trivial functions or branches in the code that are not being hit by tests at all. Goal seeking a metric can actually be harmful since it creates more test code without actually improving correctness. test code is code -- it has to be maintained too! See my comment on bitcoindevkit/bdk#952.