Warren Weckesser

A couple notes: * `mpmath` also has an implementation of `powm1(x, y)`, so this is not an unusual or highly specialized function. If both Boost and `mpmath` have it, it...

Closing for an overhaul.

I have updated the PR and marked it as ready for review. The one test failure is the lint check in the file `_add_newdocs.py` because of the indentation. I followed...

I just noticed some issues with edge cases and function domains that I need to address, so I'm marking this as *draft* again while I work out the details.

I simplified the PR, so now the two changes are: * add the new ufunc `scipy.special.powm1`, * update the existing ufunc `scipy.special.erfinv`. `gamma_ratio` and `gamma_delta_ratio` can be added in another...

Merged main to resolve a conflict. The CircleCI failure has occurred in at least one other recent pull request, and is not related to the changes here.

+1 for consistency. If we adopt sentence case as the preferred style, eventually that convention should be used consistently throughout the NumPy documentation, including in the section headings of the...

@rebeccafair, this is a shot in the dark, but maybe worth checking out. Over in https://github.com/pace-neutrons/Euphonic/blob/master/c/dyn_mat.c, in the function `diagonalise_dyn_mat_zheevd`, you have this C code: ``` int lwork, lrwork, liwork...

Is there any way for the repr of a bytes or Unicode scalar to include the actual size?

Good point, thanks. One can create instances with padded zeros directly: ``` In [23]: s = np.str_('abc\0\0') In [24]: s.dtype Out[24]: dtype('