Michael Droettboom

I have no objections if you wanted to make a pull request -- but I don't know too much about these things.

Can you merge in main to get [this fix in](https://github.com/python/cpython/pull/125477)? That way we can benchmark on Windows as well.

> Awesome! @mdboom, I'm landing this, so we should probably also update the benchmark runners that need it. Let me know if you want help. Yep -- I'll do that...

I have no comment one way or the other about these changes, but we should rename the benchmark to `pathlib2` if we do this so that comparisons won't be misleading.

> @mdboom could you help me understand what you mean by "so that comparisons won't be misleading"? When a benchmark is changed significantly like this, previously run baseline benchmarks are...

Though since not all benchmark repositories are necessary git, we should do this as part of the repository plugin and only the git one would do this.

> I'm not sure if this make sense anymore. Maybe just drop the "queued" stat? I leave that to you, but this stat is new as of #116206 (it's just...

Seems like this one is worth benchmarking (even though our gc benchmarks are a bit non-sensical, we may some things in the other ones). I fired off a run.

This PR overall improves speed by [27% GM / 0% HPT on the benchmarking suite and uses 6% more memory](https://github.com/faster-cpython/benchmarking-public/tree/main/results/bm-20240321-3.13.0a5+-d0672a4), however, this is after the regression in #116206 of [24%...

Given #117123, and the performance regression, maybe it makes sense to revert #116206 while working on getting this right?