rsmith added a comment. Thanks!
================ Comment at: include/clang/Basic/DiagnosticSemaKinds.td:4091 "%select{rvalue|lvalue}1">; +def ext_pointer_to_const_ref_member_on_rvalue : ExtWarn< + "pointer-to-member to const& function on an rvalue is a C++2a extension">, ---------------- I would be inclined to use an `Extension` rather than an `ExtWarn` for this. (I really don't see why we didn't move this as a DR -- it seems to be obviously a defect rather than an extension to me.) ================ Comment at: include/clang/Basic/DiagnosticSemaKinds.td:4092 +def ext_pointer_to_const_ref_member_on_rvalue : ExtWarn< + "pointer-to-member to const& function on an rvalue is a C++2a extension">, + InGroup<CXX20>; ---------------- Hmm, this seems hard to express clearly and concisely. How about something like: "invoking a pointer to a 'const &' member function on an rvalue is a C++2a extension" ================ Comment at: include/clang/Basic/DiagnosticSemaKinds.td:4093 + "pointer-to-member to const& function on an rvalue is a C++2a extension">, + InGroup<CXX20>; +def warn_cxx17_compat_pointer_to_const_ref_member_on_rvalue : Warning< ---------------- Please mark this as `SFINAEFailure` too; if we're going to reject under `-pedantic-errors`, we should also treat it as SFINAEable. https://reviews.llvm.org/D36855 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits