Hi! The following testcase is incorrectly rejected, because since r15-6744 it adds build_nop to preserve the exact typedef type and cxx_eval_constant_expression can't fold NOP_EXPR from integer_zerop of one NULLPTR_TYPE to another NULLPTR_TYPE, while cxx_eval_constant_expression relies on fold to do such folding.
I see 3 options to fix that, one is deal with this in the r15-6744 spot and special case NULLPTR_TYPE there and build_zero_cst instead of build_nop (with similar condition like in the patch below), another possibility is what this patch does, special case this in cxx_eval_constant_expression, and another one would be to handle this somewhere in fold-const.cc (e.g. in fold_convert_loc or there and also in fold_convert_const). Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk? Or do you prefer some other spot like mentioned above, and/or multiple spots? 2026-01-28 Jakub Jelinek <[email protected]> PR c++/123790 * constexpr.cc (cxx_eval_constant_expression) <case NOP_EXPR>: Return build_zero_cst (type) for NULLPTR_TYPE_P (type) && integer_zeropp (op). * g++.dg/cpp0x/nullptr47.C: New test. --- gcc/cp/constexpr.cc.jj 2026-01-28 09:34:32.746243466 +0100 +++ gcc/cp/constexpr.cc 2026-01-28 19:00:14.201290765 +0100 @@ -10138,6 +10138,10 @@ cxx_eval_constant_expression (const cons break; } + /* fold doesn't optimize NULLPTR_TYPE casts. */ + if (NULLPTR_TYPE_P (type) && integer_zerop (op)) + return build_zero_cst (type); + if (op == oldop && tcode != UNARY_PLUS_EXPR) /* We didn't fold at the top so we could check for ptr-int conversion. */ --- gcc/testsuite/g++.dg/cpp0x/nullptr47.C.jj 2026-01-28 19:03:10.860279799 +0100 +++ gcc/testsuite/g++.dg/cpp0x/nullptr47.C 2026-01-28 19:02:48.805655698 +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
