chore: bump python version used for codecov in CI #520
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #520 +/- ##
=========================================
Coverage 100.00% 100.00%
=========================================
Files 22 22
Lines 2096 2096
Branches 347 291 -56
=========================================
Hits 2096 2096 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
| @@ -69,7 +69,7 @@ jobs: | |||
| run: | | |||
| pytest --runslow | |||
| - name: Upload coverage to codecov | |||
| if: matrix.python-version == '3.8' | |||
| if: matrix.python-version == '3.9' | |||
There was a problem hiding this comment.
Currently the coverage only runs for 3.8 (and other versions run pytest without it), so this needs to change in the two places above as well. Have a look at #428, possibly this update means we can also remove the typing_extensions install there.
There was a problem hiding this comment.
but this
- name: Test with pytest
if: matrix.python-version != '3.8'
run: |
pytest --runslow
stays the same, right? I am guessing the reason to only run the slow tests on py>3.8 had a reason beyond 3.8 being the oldest version supported.
There was a problem hiding this comment.
No, I think they should all change. The logic is:
- run pytest with coverage for a specific version (previously 3.8, now 3.9),
- run pytest without coverage for the other versions (previously !=3.8, now !=3.9),
- upload if on the right version (previously 3.8, now 3.9).
I'm not sure what the 3.9 CI actually uploaded now because it didn't seem to fail. That surprised me since it should not have generated a coverage report in the current state.
There was a problem hiding this comment.
right but then for #428 to be dropped we would need to drop 3.8 all together, because the CI will still be running that (why it currently fails)
There was a problem hiding this comment.
right, I see now it's unrelated to coverage
In #515 the coverage was observed to be underestimated due to the use of
xfail()to handle the issue described in #476. This reduced coverage was not observed locally (withpython 3.9.7). It happens because the CI still usespython 3.8to report coverage. This python version will be eventually deprecated, and it makes sense to not use it for code coverage anymore.