AFAIK, the idea back then was that single/multiple phase should be determined by the return type from the init function, not by the existence of slots; NULL slots would be equivalent to zero slots (a multi-phase module that doesn't need custom initialization).

I see that multiphase with m_slots=NULL isn't tested explicitly, though.

Thanks for bringing this up. What you've said makes sense and appreciate the clarity of it. There are indeed some gaps (fairly small, I expect), both functionally and in tests. It isn't clear what the impact is in practice, but I definitely think they should be addressed regardless. I'll be sure to open some issues in the next week or two.

FWIW, the general thought of kinds of extension modules was already on my mind (and a bit clearer in my original mega-PR, gh-118116). For now I have a later PR (gh-118205) that is a bit more deliberate about keeping track of what the init function returns. While that PR doesn't do so currently, I might look at explicitly tracking the kind on the module def. That would help address the point you've brought up.

Thanks! Following up here: #117953 (comment)