Matt Haberland

Note that this failure only occured on our [macos_arm64_test](https://github.com/scipy/scipy/runs/10962045507). This is the same platform in which we see the overflow warnings. Actually, they could be related. I didn't realize it...

> Our goal for quantile is to find integer k such that negative_binomial(k, n, p) >= probability In general, the PPF is the *smallest* integer k such that the CDF...

I think the situation is improved by terminating early, yup. Less likely to run into the trouble you described. This is making me wonder about how > one end randomly...

> But it's passing the tests, what am I missing? Maybe the architecture? https://github.com/boostorg/math/issues/935#issuecomment-1409445468

I don't. Maybe @mckib2 can help with that Could you try updating the termination tolerance, though? It seems like it it's good for performance whether or not it fixed this...

Yeah I actually did run into this special case while converting code from NumPy to array API. The purpose of the code is to produce a different status flag that...

Just thought I'd mention that I ran into this again in a different context.

Re: https://github.com/HDembinski/jacobi/issues/37#issue-1774280091, it depends on the intent of relative `x`. IIUC, they should only be helpful for very large `x`, when you are losing precision in the abscissae at which...

Re: https://github.com/HDembinski/jacobi/issues/37#issuecomment-1607117562 , that makes sense, but then you lose speed when one-sided differences aren't needed at all. I like the current approach - start with central difference, see whether...

Ah, right. Personally, i think it is still enough for automatic detection, though. You are making an effort to detect an obvious case where central difference is invalid, but you...