GitHub Actions: Upgrade to codecov v4

[cclauss]

Short description

GitHub Actions: Upgrade to codecov v4

Proposed changes

• https://github.com/codecov/codecov-action?tab=readme-ov-file#example-workflowyml-with-codecov-action

Resolved issues

Fixes: #69

[timobrembeck]

@cclauss I really appreciate your effort, but if it doesn't work out in a reasonable amount of time, I think it might not be worth it. Maybe it's really a bug in the upstream action which will be resolved in the future. For the moment, v3 should work just fine for us 🥲

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.64%. Comparing base (e9d4308) to head (8339a13).

Additional details and impacted files

@@             Coverage Diff             @@
##              main      #70      +/-   ##
===========================================
- Coverage   100.00%   99.64%   -0.36%     
===========================================
  Files           10       20      +10     
  Lines          328      559     +231     
===========================================
+ Hits           328      557     +229     
- Misses           0        2       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[cclauss]

Snap!

[timobrembeck]

Awesome, thanks for figuring that out!
Maybe we have to exclude setup.py explicitly from the coverage report to not get a decrease in coverage?
In my opinion, this is justified, since the setup itself is usually never covered by the unit tests, right?

[cclauss]

setup.py usually goes untested and more and more it is just getting removed.

• https://setuptools.pypa.io/en/latest/deprecated/commands.html

[timobrembeck]

setup.py usually goes untested and more and more it is just getting removed.

* https://setuptools.pypa.io/en/latest/deprecated/commands.html

Right, I'm also fine with removing it. 👍
It should have no downsides for users installing this package via pip, right?

[cclauss]

Let's not lose setup.py in this pull request. I doubt the loss of one-third of 1% is caused by the presence of setup.py. Once this PR is merged then 99.64% will be the coverage's new normal (green). See section 06 of https://about.codecov.io/resource/what-is-code-coverage on why 100% coverage is not vital.