Henry Schreiner

I'm guessing this is https://github.com/pybind/pybind11/issues/4105.

I think we all had decided to support this some time ago, it was just hard to do with many inflight changes also happening. Now we are are down to...

I had dual conferences last week (a problem with virtual conferences), and I just started teaching a class. Once I get a bit more caught up on writing material, and...

> permanent tax on development velocity I don't know if this is true. Splitting between header and source isn't necessarily that much of a tax - it's how C++ was...

The main thing I'd like to ensure is that we can support C++20/23 modules well, but I'd assume that's easier or similar with this change.

Not me, but the regressions others found I listed in https://github.com/pybind/pybind11/issues/4125. The regression for me has been the scikit_build_example one, which I need to work out before a release. The...

The main issue is: ``` /__w/pybind11/pybind11/include/pybind11/stl_bind.h:91:7: error: no matching function for call to 'pybind11::class_::def(pybind11::detail::op_)' cl.def(self == self); ``` I think the other issue might be new to 11.7.

Wheels don't include the tests, yes. Do you see something you could improve, like "starting with a git checkout" or something like that?

What DLL? What OS? Could you paste the error? We shouldn't be requiring a DLL, this is a header only no dependency project...

I have a feeling it has something to do with https://docs.python.org/3/library/os.html#os.add_dll_directory (new in Python 3.8) - for security reasons, Python has to be told where modules live. If you are...