On Thu, Jan 29, 2026 at 06:59:34PM +0800, Jason Merrill wrote: > I guess since NULLPTR_TYPE is in tree.def it makes sense for > fold_convert_loc to handle it.
Here is a fold-const.cc version of that. Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? 2026-01-30 Jakub Jelinek <[email protected]> PR c++/123790 * fold-const.cc (fold_convert_const): Handle conversion of integer_zerop to NULLPTR_TYPE. (fold_convert_loc): Likewise. * g++.dg/cpp0x/nullptr47.C: New test. --- gcc/fold-const.cc.jj 2026-01-14 13:21:06.136656098 +0100 +++ gcc/fold-const.cc 2026-01-29 12:48:18.263320685 +0100 @@ -2566,6 +2566,8 @@ fold_convert_const (enum tree_code code, return v.build (); } } + else if (TREE_CODE (type) == NULLPTR_TYPE && integer_zerop (arg1)) + return build_zero_cst (type); return NULL_TREE; } @@ -2788,6 +2790,10 @@ fold_convert_loc (location_t loc, tree t tem = fold_ignored_result (arg); return fold_build1_loc (loc, NOP_EXPR, type, tem); + case NULLPTR_TYPE: + if (integer_zerop (arg)) + return build_zero_cst (type); + /* FALLTHRU */ default: if (TYPE_MAIN_VARIANT (type) == TYPE_MAIN_VARIANT (orig)) return fold_build1_loc (loc, NOP_EXPR, type, arg); --- gcc/testsuite/g++.dg/cpp0x/nullptr47.C.jj 2026-01-29 12:43:59.155035248 +0100 +++ gcc/testsuite/g++.dg/cpp0x/nullptr47.C 2026-01-29 12:43:59.155035248 +0100 @@ -0,0 +1,10 @@ +// PR c++/123790 +// { dg-do compile { target c++11 } } + +int +main () +{ + using nullptr_t = decltype (nullptr); + constexpr nullptr_t zero = nullptr; + constexpr nullptr_t other_zero = zero; +} Jakub
