Closed lillian542 closed 3 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 97.98%. Comparing base (
8c191be
) to head (764059b
). Report is 200 commits behind head on main.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
Hello. You may have forgotten to update the changelog!
Please edit doc/changelog.md
on your branch with:
Context:
In
test_measurements_results.py
, we explicitly enable and then disablenew_opmath
inside each test. Enablingnew_opmath
is no longer necessary, because it's the default. Disabling it at the end of each test means that when the test is completed, we are set to use legacyopmath
.Locally we run each test file separately, but in CI, any tests on the same runner that are run after this file are run with new_opmath disabled. This can lead to discrepancies between local and CI testing.
Description of the Change: We remove the explicit enabling and disabling of
new_opmath
Benefits: We will be testing everything with new opmath in CI as intended, and local and CI tests will correspond better to each other.
Possible Drawbacks: This came up in a couple of PRs because added functionality failed with legacy opmath (otherwise this would never have been visible). Without this, we wouldn't have caught the legacy opmath incompatibility. But if we want to check for incompatibility with legacy opmath, we should add a workflow to the test matrix checking the whole thing regularly.