Closed adiroiban closed 2 months ago
Diff from pydoctor_primer, showing the effect of this PR on open source code:
twisted (https://github.com/twisted/twisted): typechecking got 1.07x slower (231.7s -> 246.9s)
(Performance measurements are based on a single noisy sample)
According to pydoctor_primer, this change doesn't affect pydoctor warnings on a corpus of open source code. ✅
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 92.67%. Comparing base (
7e0a09f
) to head (ba0c62f
).
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
According to pydoctor_primer, this change doesn't affect pydoctor warnings on a corpus of open source code. ✅
According to pydoctor_primer, this change doesn't affect pydoctor warnings on a corpus of open source code. ✅
please review
Fixes #782
This removes the tox helper to publish the coverage.
It used the dedicated github action from Codecov. The never version of codecov uploader need an explicit token.
Also, if codecov upload fails, the job fails. In this way it should be easier to detect coverage reporting errors.
As a drive by, I have also updated some other GitHub Actions to reduce the warnings during build.
As a drive by, I have also removed some trailing spaces.
In a separate ticket/PR we can look at running the coverate report using only GItHub Actions, similar to what we have for twisted/towncrier ...but use diff_cover
https://github.com/twisted/towncrier/blob/914b446e64d62f3968a97931e490b1abe5c8c581/.github/workflows/ci.yml#L258-L267