On Fri, Apr 14, 2023 at 5:56 PM Stefan van der Walt <stefanv@berkeley.edu> wrote:
Hi Ralf,

On Fri, Apr 14, 2023, at 08:57, Ralf Gommers wrote:
- you have to (unfortunately) also still deal with the setup.py-based build in one way or the other. We're going to throw it out soon, so you can either ignore it and keep the slow `_mmio.py` implementation for that, or integrate the C++ code in `scipy/io/setup.py`. The latter is a little nicer, but in case it's tricky to get to work then keeping the slow code is also okay. The only reason we will still keep it around for the 1.11.0 release (probably) is conda-forge on Windows, so that's a limited number of users who'd get the slow code.

I'd be curious to learn more about the build issues on conda-forge; is there a related issue? skimage just did a cross-compiled conda-forge release using the new meson machinery, but of course it's a significantly simpler setup than SciPy.

Oh that's the usual "there is no suitable Fortran compiler on Windows". See https://github.com/conda-forge/scipy-feedstock/issues/213 for details. We have a custom Mingw-w64 based toolchain for our own wheels, but that doesn't translate too well to any other packaging ecosystem where the compiler itself is distributed as a package and things have to be built in as uniform a way as possible.

Cheers,
Ralf