erichkeane wrote: > This attempts to fix #26078. However I have couple of fundamental questions > with the 7.1.5 consexpr note "[ Note: An explicit specialization can differ > from the template declaration with respect to the constexpr specifier. — end > note ]"
Current quote: https://eel.is/c++draft/dcl.constexpr#note-1 is the same. > > 1. Does it also apply to member specializations which are not function > templates themselves ? (as in the example https://godbolt.org/z/c4bYY8rxb) I read that to do so, yes. It applies to ALL templates. An explicit specialization can change the constexpr/constevalness. > > 2. if yes, then should it also apply to special member functions ? > (constructors, destructors) I would say yes, it should. Any templated SMFs also can have it changed with an explicit specialization. This is actually related to a topic that came up for Reflection last week in EWG. An 'explicit specialization' is a morally different 'entity'. So it is allowed to change quite a bit more from the primary template than we might otherwise expect. In this case, it is allowed to add/remove constexpr/consteval. https://github.com/llvm/llvm-project/pull/145272 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
