Henry Schreiner

Is this with the classic discovery or the new discovery? It seems you are setting both `Python*` and `PYTHON*` variables. And `PYTHON_MODULE_EXTENSION` was not supposed to be overridable, it is...

Okay, I followed links and it looks like it's partially due to a bug in pypy37. :(. We could probably make this something that can be user overridable. We could...

(And 2.6.2 was before #3299 which is where the rediscovery support was added)

Hmmm, CMake failing to find PyPy 3.9 but reporting 3.9 is a valid version was not what I was expecting. That works correctly locally.

This is the local error: ``` nox > cmake --build .nox/pypy_upstream-3-9/tmp --target pytest [0/2] Re-checking globbed directories... [0/1] cd /Users/henryschreiner/git/software/pybind11/.nox/pypy_upstream-3-9/tmp/tests && /Users...tests/test_union.py /Users/henryschreiner/git/software/pybind11/tests/test_virtual_functions.py libc++abi: terminating with uncaught exception of type...

With that fix, Python tests pass, then the CMake tests fails with: ``` ======== CMake output ====== Found pybind11: /Users/henryschreiner/git/software/pybind11/.nox/pypy_upstream-3-9/tmp/mock_install/include (found version "2.10.0dev1") Found pybind11 v2.10.0 dev1: /Users/henryschreiner/git/software/pybind11/.nox/pypy_upstream-3-9/tmp/mock_install/include;/Users/henryschreiner/git/software/pybind11/.nox/pypy_upstream-3-9/tmp/pypy-c-jit-104891-17e3c4e34398-osx64/include/pypy3.9/ Configuring done...

3.7 and 3.8 were always working. It's just 3.9 that's broken - via CMake on CI and by `_PyPyCMethod_New` missing (have no idea what that is) locally - only in...

Yes, but it is finding it for 3.8 & 3.7, AND it is finding it locally on my Mac for 3.9, even though everything should be identical - same pypy...

The normal tests pass, then the build tests fail with the error listed above.

Let me revert the last commit, that should get past that test.