Not really a support request, more like a general FAQ-like question:
So a bit of background: I am one of the maintainers of the
aiosmtpd library. We are currently transitioning from Travis CI to GitHub Actions (GA). One of the steps in the GA Workflow we’ve made is the uploading of the coverage testing result to codecov.io.
Now since this is a FOSS library recommended for use by Python documentation, naturally we have to test it against several Python versions & implementations (currently 3.6 to 3.9, inclusive, and PyPy3). We implement the testing using
tox which sets things up and perform the actual unit testing & coverage testing (using pytest & pytest-cov).
This resulted in several
coverage*.xml files, one per each Python version/implementation.
So my questions:
- Should we upload all of the XML files (in parallel through several GA jobs) or just one?
- If we upload all of them, what will happen with the coverage report? Will Codecov merge them additively, or provide separate reports based on the version/implementation, or both? Or?
- Should I somehow tag/flag/mark the uploads separately? And if so, how?
Thank you in advance.