Fix failing CI

[di]

CodeCov's services seem to be getting more and more unavailable, and are causing our tests to fail (e.g., https://github.com/pypa/twine/actions/runs/3577489389/jobs/6016720780).

This PR removes CodeCov entirely in favor of running coverage in CI and asserting on some coverage percentage, currently 97%.

Fixes #658. This includes #942 to get CI to pass, closes #942. Closes #936.

[di]

Looks like no_color for rich is not getting picked up on Windows now for some reason?

>       assert capsys.readouterr().out == f"Checking {sdist}: PASSED\n"
E       AssertionError: assert 'Checking C:\....gz: PASSED\n' == 'Checking C:\....gz: PASSED\n'
E         Skipping 64 identical leading characters in diff, use -v to show
E         - wn\pytest-0\test_passes_markdown_descripti1\dist\test-package-0.0.1.tar.gz: PASSED
E         + wn\pytest-0\test_passes_markdown_descripti1\dist\test-package-0.0.1.tar.gz: PASSED
E         ?           ++++ ++++                                                   ++++   ++++ ++++  ++++

tests\test_check.py:270: AssertionError

[bhrutledge]

Looks like no_color for rich is not getting picked up on Windows now for some reason?

Yeah, I started looking into this in #936, but haven't had the time to dig into it.

[sigmavirus24]

We still care about coverage and there are ways to display it in CI. But we can add it back eventually I guess

[di]

We still care about coverage and there are ways to display it in CI. But we can add it back eventually I guess

The results are still displayed in CI: https://github.com/pypa/twine/actions/runs/3577925730/jobs/6017490264#step:6:110

Do you mean adding comments on PRs like CodeCov did?

[bhrutledge]

The results are still displayed in CI: https://github.com/pypa/twine/actions/runs/3577925730/jobs/6017490264#step:6:110

As noted in #658 (comment) (and subsequent comments), I think it'd be nice to make the HTML report available.

[di]

I think once nedbat/coveragepy#1418 is released, this would be as simple as:

- python -m coverage report -m --fail-under 97
+ python -m coverage report -m --fail-under 97 --format=markdown >> $GITHUB_STEP_SUMMARY

https://github.blog/2022-05-09-supercharging-github-actions-with-job-summaries/

[webknjaz]

CodeCov's services seem to be getting more and more unavailable, and are causing our tests to fail

FWIW the workflow used a long-deprecated version of the action (v1). The current v3 is usually working well.

[webknjaz]

once nedbat/coveragepy#1418 is released

@di FYI, it is possible to implement this right now with coverage xml, irongut/CodeCoverageSummary action and a few lines of code. Here's how I did it: https://github.com/ansible-community/ansible-test-gh-action/blob/ce3e835/action.yml#L493-L511.