Closed konstin closed 2 months ago
Pubgrub is so fast to compile that caching in CI isn't required, so i removed it. Closes #201.
I've added workflow_dispatch to dry run these workflows.
workflow_dispatch
The timing below are from my machine, but even if CI is an order of magnitude slower i don't think we should cache:
$ hyperfine --prepare "cargo clean" "cargo publish --dry-run" "cargo +nightly doc --no-deps" Benchmark 1: cargo publish --dry-run Time (mean ± σ): 1.661 s ± 0.272 s [User: 2.240 s, System: 0.324 s] Range (min … max): 1.467 s … 2.207 s 10 runs Benchmark 2: cargo +nightly doc --no-deps Time (mean ± σ): 1.230 s ± 0.020 s [User: 2.013 s, System: 0.312 s] Range (min … max): 1.202 s … 1.263 s 10 runs
Looks good to me, but why the lock changes in this PR? It seems unrelated.
Not sure what was out of sync, but solved by rebasing on dev
Pubgrub is so fast to compile that caching in CI isn't required, so i removed it. Closes #201.
I've added
workflow_dispatch
to dry run these workflows.The timing below are from my machine, but even if CI is an order of magnitude slower i don't think we should cache: