Albert Steppi

Thanks for the bug report @f-brinkmann. At a glance, I would guess this is due to logic for handling the branch-cut in the general case which shouldn't be applied in...

> [@steppi](https://github.com/steppi) could you already find time to look into this? If it's of any help, I could offer to write tests for this. I initially stumbled upon this, while...

Hi @mberz. I haven't had bandwidth to look at this yet. My time situation is improving though and I'll try to take a look this week.

> [@steppi](https://github.com/steppi) Thank you very much. If you cannot find the time, you can also point us in the right (or rough) direction and I can see if I can...

Interesting. I suspect this fails for all complex valued function in `cython_special`. My guess is that the issue relates to the type punning code that's used to convert between `npy_cdouble`...

> Actually, upgrading to NumPy 2 is a reasonable resolution of the issue for debian. Its packaging is under way, it'll be available for debian before too long. We can...

> @steppi would interpolation be ok, or were you hoping for the definition to match that of evaluation in terms of `binom`? Yeah, I was hoping this would evaluate $$\frac{{r...

Are you OK with decoupling the rename from the improvements? I'd like to see a rename of opaquely named cdflib functions too, but would prefer that a naming scheme be...

> I will merge this within a week if there are no further comments. Boost is clearly better and this is a rather niche function. I finally have a chance...

https://github.com/scipy/xsf/pull/39 landed in `xsf` and I've updated this PR to use the latest `scipy/xsf` release in the submodule instead of a branch on my fork. I want to make sure...