Matt Haberland

Is this all set @Kai-Striega ? Want us to look again or are there still unresolved comments or other changes?

@Kai-Striega was this ready for another look or were you still working on those unresolved comments?

I'd prefer to wait until those are resolved before taking a final pass.

Thanks for this. Speedup does look impressive! Can't help with the failing build - looks like it might be real, as `stats/setup.py` and Pythran are mentioned in the log. I...

Status update: `_compute_prob_inside_method ` no longer exists, so that part of this PR is no longer needed. `%timeit _compute_prob_outside_square(3000, 676)` The original Python version: 740 µs ± 1.36 µs per...

Since I already merged main into this one, I didn't want to do it again for gh-14314, so let's bring `_count_paths_outside_method` over here. For that, the speedup is much more...

Tests pass locally, but I know I'll need to add a new test because: ```python3 import numpy as np from scipy import stats # one-sided, n1!=n2 -> _count_paths_outside_method rng =...

@serge-sans-paille I saw that you supported some of these Pythran/SciPy PRs before, so I wonder if you can help me finish this one? There are overflows happening inside `_count_paths_outside_method`. Previously,...

@serge-sans-paille thought I'd check in again on whether you can support this, or if you can suggest someone else who might help on the Pythran side. Thanks for considering it!

Thanks @serge-sans-paille! The snippet [here](https://github.com/scipy/scipy/pull/13957#issuecomment-1200357297) will trigger the error because an overflow occurs in `_count_paths_outside_method`. --- In case it helps, here's the context. https://github.com/scipy/scipy/blob/07fc190d7a3aeed421f266765eb150bea89bf21a/scipy/stats/_stats_py.py#L7836-L7837 If you run the same snippet...