Use 'cargo insta test' in Makefile's test command

[cnpryer]

I'm sure it's intentional to not run everything in CI. I felt it might be helpful especially for the CLI to fully run.

[cnpryer]

Plucked from https://github.com/astral-sh/ruff/blob/fc8738f52af124ab54df61c7a158bf46fe8e7532/.github/workflows/ci.yaml#L121.

I haven't seen this yet https://github.com/taiki-e/install-action.

[konstin]

Have you considered using nextest? For rye's small test suite the difference isn't big yet, but for uv it makes difference.

$ hyperfine --runs 5 "cargo nextest run" "cargo test"
Benchmark 1: cargo nextest run
  Time (mean ± σ):     11.164 s ±  0.091 s    [User: 7.781 s, System: 2.562 s]
  Range (min … max):   11.088 s … 11.313 s    5 runs
 
Benchmark 2: cargo test
  Time (mean ± σ):     17.686 s ±  0.532 s    [User: 7.825 s, System: 2.837 s]
  Range (min … max):   16.968 s … 18.447 s    5 runs
 
Summary
  cargo nextest run ran
    1.58 ± 0.05 times faster than cargo test

[cnpryer]

Have you considered using nextest? For rye's small test suite the difference isn't big yet, but for uv it makes difference.

I haven't used nextest yet personally, although I've known of it. I've been playing around a little bit with the uv tests just to look around, and I've liked nextest so far.

I'd be happy to check that out for this PR.

[mitsuhiko]

I would prefer that the CI invokes a makefile command. Other than that we can try this.

[cnpryer]

I would prefer that the CI invokes a makefile command. Other than that we can try this.

Okay, I'll revert back to what I had then. I'll push that real quick.

If you decide to merge before nextest I'll PR that separately soon. I don't anticipate getting to the nextest PR before wrapping up #759.