Closed devmotion closed 3 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 85.96%. Comparing base (
f852803
) to head (8aee234
).
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
without running tests twice if no prerelease is available.
I don't think that will happen. There will only be a single run for julia@1
but that run might be a prerelease, see https://github.com/JuliaStats/Distributions.jl/actions/runs/9250768369/job/25444947763?pr=1862, and maybe that is not what we want.
There will only be a single run for julia@1 but that run might be a prerelease,
Indeed, my comment above was incorrect: I was already assuming that we would want to change that PR such that we continue testing with the latest stable release.
I'm wondering if it's worthwhile to add all the complexity to the CI.yml file for this, though. I think testing prereleases in dedicated PRs is reasonable. It requires a little effort but not much.
I agree, in particular given that tests seem to fail on the prerelease as well (and the fact that the PR in its current state actually doesn't skip the duplicate runs).
In this draft I see if one can achieve something like #1861 without running tests twice if no prerelease is available. Additionally,This PR removes the nightly tests (see #1861), includes a fix for the mac CI setup (see #1856), and switches to julia-actions/cache for caching.Edit: The PR also includes https://github.com/JuliaStats/Distributions.jl/pull/1862 which fixes tests on Julia 1.11-beta1 (see https://github.com/JuliaStats/Distributions.jl/pull/1862#issuecomment-2133916510).