https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119620
--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to 康桓瑋 from comment #0) > I don't know why the standard sometimes constrains emplace(), sometimes only > constrains insert(), and sometimes constrains neither. flat_multiset::emplace is specified to construct an element explicitly, which requires is_constructible_v<value_type, Args...> *in addition to* being able to emplace that object into the container as shown in the Effects paragraph. flat_set::emplace is not specified to unconditionally construct an object outside the container (even though for some calls to emplace it does need to do that). It seems like we're missing a precondition on flat_set::emplace, or that flat_multiset::emplace could be specified differently so that it doesn't need to construct an element outside the container.
