How would the PyPA say that? That’s a serious question - I genuinely don’t know how the average user distinguishes between a formal recommendation from the PyPA and a bunch of random documentation they found on the internet. Is packaging.python.org genuinely that influential?
I say that very much aware that if you take packaging.python.org as definitive, the PyPA recommendation is already to use hatch. It’s a pretty mild recommendation, because there was a lot of agonising when we did the rewrite about not supporting people who chose other tools, or repeating the fiasco of pipenv, but it’s definitely there. And yet the survey results suggest that the vast majority of users aren’t following that recommendation.
As regards “clear documentation”, we can write documents on packaging.python.org, but ultimately a lot of it is going to be down to the tool maintainer(s), not the PyPA.
Yeah. So setuptools, then. People are working on alternatives (although they are mostly working on the hard problems in that area, not so much on simpler “I wrote an accelerator in C” use cases) but right now, native code means setuptools, and unpleasant problems like fighting with gcc vs MSVC on Windows.
We simply don’t have a better answer for native code extensions yet. So does that mean we offer no answer to the user complaints about complexity, or do we offer an answer now, with the qualification that it doesn’t cover native code very well, so for that you need to fall back on other, well-supported and stable but not “unified” solutions?
Again, that’s a genuine question that I don’t know the answer to. The survey didn’t (as far as I know) break respondents down by whether they care about native code, so we have no information there. Nor did it distinguish between people who are happy to just consume prebuilt binaries, and people who want (need) to build from source.