Henry Schreiner

You are welcome to take `PYBIND11_EXPORT_EXCEPTION` concerns very seriously; we can release a 2.10.2 if we come up with a further fix. This is a situation that is so complex...

> Why slip into such radical language? I'm just quoting you. You have said we need to abandon the 2.10 release series. Multiple times (https://github.com/pybind/pybind11/pull/4218#issuecomment-1287307680, https://github.com/pybind/pybind11/issues/4125#issuecomment-1273623503). If we could "yank"...

"Abandon a release series to protect users" is abandoning users. If users have already moved to 2.10 (as many of them have), refusing to release important bug fixes because you...

Have you tested the latest master of pybind11? We fixed a variety of things that could possibly be related. Have you tested building the parent library with 2.10? IMO, this...

> resulting in CMake building Python extensions that did not match the Python interpreter version I believe this was fixed with #3577, which would have been in 3.10.1. This is...

I verified this is not #4105, this code was broken in 2.9 as well.

To be fair, there was a method doing _exactly_ this in the original particle; it was called `__int__`. This method would have solved the numba issue, but it was removed,...

> What is the problem with inheriting from PDGID? These were developed separately (I wrote Particle, Eduardo wrote PDGID). Though, I do see it as a separation of concerns. PDGID...

Is a Particle an int? I think no, it is not an int. There is an int associated with it, yes, but it _is_ not an int. You can have...

Let's get #527 in, a 0.8.1 out, and then go for this.