https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120259
--- Comment #3 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Tomasz Kamiński from comment #2) > This seems to be regeressions from PR1579: PR71579 > r11-2759-g69f571ffc513b689fa26e4c9fceba17c2c989ab3 > r11-3442-gc1fc9f6e10e646f01194c8f150affbc1cfbc404a > > I think we went to eager there, and most likely should revert. I thought we were required to diagnose these, which is what was proposed for LWG 2797, but what was actually approved in https://wg21.link/p1285r0 just says it's undefined. So we don't have to diagnose it if doing so is a problem. I agree we should revert the static_assert for the swappable traits. For a type where swapping uses the generic std::swap, we should still get diagnostics for incomplete types from the is_move_constructible and is_move_assignable traits.
