Andrew Nelson

https://github.com/scipy/scipy/runs/7854253531?check_suite_focus=true shows that the win_amd64 DLL aren't importable on 3.10 with this version, 1.22.3 works.

https://github.com/scipy/scipy/pull/16842/commits/c8047d926d9c819aae3734854ad7bdb8f2a830de#diff-50c86b7ed8ac2cf95bd48334961bf0530cdc77b5a56f852c5c61b89d735fd711R41 `"numpy==1.22.3; python_version=='3.10' and platform_system=='Windows' and platform_python_implementation != 'PyPy'"` seems to work. i.e. `platform_machine-->platform_system`

Some docstrings have multiple examples, but these are often interspersed with lots of text. Should a numpy import happen once in a docstring, or with each example (as previous comments...

Plus it's difficult for a lot of devs to make meetings due to time zone issues.

I'm with Ralf on the dependencies, I definitely wouldn't want to be adding more hard dependencies, only optional ones (if at all). In any case, it's probably worth asking ourselves...

I don't think it matters which CI one uses in terms of speed. My initial thought is for GHA, simply because you can have a standalone yml config. However, if...

I don't know what the scipy plan is to handle that transition. i.e. when do we need to make a release that only builds against numpy2.0 and newer. I'd have...

e.g. at the moment the scipy pyproject.toml file will want to build against `numpy==1.22.4` for Python 3.10 wheels, that's what the last scipy would've been built against. If we start...

@jakirkham yes, the scipy nightlies are being built against `numpy>=2.0.0.dev0`. They are available from https://anaconda.org/scientific-python-nightly-wheels/scipy/files.

There are regular nightly builds which should aid in prerelease testing