I’m a maintainer on cibuildwheel, along with @yannickjadoul, @mayeut and @henryiii. If you haven’t come across it yet, cibuildwheel is a tool for Python wheel building that makes building platform wheels in CI simple - across all the OSs and Pythons that we support.
The project is reaching a pretty stable state, and appears to be gaining in adoption across the community (notable users include mypy, scikit-learn, matplotlib, coverage.py…) so we’ve been starting to think about more long-term plans, to ensure it’s still around and working for the community for years to come.
One thing we were wondering about was if the PyPA would be interested in adding cibuildwheel as a member project? cibuildwheel is already stitching together lots of pypa work, like the manylinux project, auditwheel, virtualenv, of course pip & wheel, perhaps soon pypa-build and musllinux! We end up being the interface to wheel building for many users, so perhaps it would be useful to have us more involved in discussions relating to platform wheels? From our perspective, we’d benefit from a little more visibility, and being under the pypa umbrella might give users more confidence that we’re the real deal
So - I’d love to know what PyPA members think - is cibuildwheel a good fit, do you think? Would it make sense for cibuildwheel to be a ‘1st party’ part of PyPA, or would it serve the community better as part of the ecosystem? Perhaps there are benefits to joining/staying independent that we haven’t thought of? Let us know your thoughts below.