Albert Steppi

> Let's check with @steppi re: the comment on my patch there: I have a better understanding of how the ufunc generation infrastructure works now, and @izaid and I started...

> @steppi, are you working on this currently? I can work on it, but I won't bother if you (or someone else) already have a fix in progress. I'm not...

> I think `int` will result in the same problem. My understanding is that even on 64 bit Windows, `int` is still 32 bit, but the default NumPy integer type...

> > OK, I see that [intp](https://numpy.org/doc/stable/reference/arrays.scalars.html#numpy.intp) has character code "n" and should work in a ufunc. > > Right, `n` is not available on NumPy 1.x though, so you'll...

@fbourgey, I don't understand where the updates to the `xsf` submodule come from. Did you edit the `xsf` code within SciPy and make a PR here? Updates to xsf should...

> @steppi, yes, I've updated the `legendre.h` file. Sorry, first time modifying such files in scipy. No worries, it's not your fault. The `xsf` submodule is still pretty new and...

> > For the dtype fix, I am also not clear why there is a need for the new dtype= kwarg actually. > > The reasoning for adding the `dtype`...

I've added a subsection in #24130 which contains the recommendation I made above ```rst For array creation functions without array inputs, adding array API standard support can be accomplished by...

Thanks @rgommers for weighing in. Since the intention of JAX eager mode is for debugging something which one wants to run in jit mode, do I understand correctly that no...

> > I guess this PR should be shelved? > > Please let's not be too hasty. _I_ am knee deep in trying to add support for all array backends...