[v4] c++/reflection: P4101, Consteval-only values [PR125820]

On Tue, Sep 01, 2026 at 11:20:30PM -0400, Jason Merrill wrote:
> On 9/1/26 7:47 PM, Marek Polacek wrote:
> > On Thu, Aug 27, 2026 at 02:39:30PM -0700, Jason Merrill wrote:
> > > On 8/27/26 4:45 PM, Marek Polacek wrote:
> > > > On Tue, Aug 25, 2026 at 03:11:21PM -0700, Jason Merrill wrote:
> > > > > On 8/25/26 12:56 PM, Marek Polacek wrote:
> > > > > > On Mon, Aug 24, 2026 at 03:20:33PM -0700, Jason Merrill wrote:
> > > > > > > On 8/23/26 6:08 PM, Marek Polacek wrote:
> > > > > > > > @@ -1136,6 +1136,41 @@ store_init_value (tree decl, tree init, vec<tree, va_gc>** cleanups, int flags)
> > > > > > > >        /* Handle aggregate NSDMI in non-constant initializers, too.  */
> > > > > > > >        value = replace_placeholders (value, decl);
> > > > > > > > +  /* Detect stuff like 'info r = ^^int;' outside a manifestly
> > > > > > > > +     constant-evaluated context.  */
> > > > > > > > +  if (flag_reflection
> > > > > > > > +      && !processing_template_decl
> > > > > > > > +      && !DECL_DECLARED_CONSTEXPR_P (decl))
> > > > > > > > +    {
> > > > > > > > +      bool bad = check_out_of_consteval_use (value, /*complain=*/false);
> > > > > > > > +      /* A non-constexpr variable at namespace scope with a constant
> > > > > > > > +	 initializer has constant initialization, so we check the folded
> > > > > > > > +	 value not to wrongly reject "int e = (^^int, 42);".  For non-static
> > > > > > > > +	 local variables, there is no such rule, so we check the unfolded
> > > > > > > > +	 initializer.  But we should also reject
> > > > > > > > +
> > > > > > > > +	   consteval auto fn () { return ^^int; }
> > > > > > > > +	   void g() { auto r = fn (); }
> > > > > > > > +
> > > > > > > > +	 so we may have to check both.  Note that the first call could
> > > > > > > > +	 have escalated and so we may find ourselves in an immediate
> > > > > > > > +	 context now.  */
> > > > > > > 
> > > > > > > I don't think we need to check both; we should check one or the other
> > > > > > > depending on DECL_INITIALIZED_BY_CONSTANT_EXPRESSION_P.
> > > > > > 
> > > > > > I don't know if I could do away with the second call.  I want to give
> > > > > > an error for `r` in the comment where it is DECL_INITIALIZED_BY_C_E_P
> > > > > > and I have to check value to see the REFLECT_EXPR.  But in
> > > > > > 
> > > > > >      void g() { int e = (^^int, 42); }
> > > > > > 
> > > > > > `e` is also DECL_INITIALIZED_BY_C_E_P but I have to check init
> > > > > > to find that ^^int because value is just 42.
> > > > > 
> > > > > Hmm, true, I was thinking it was only set just above for MCE initializer.
> > > > > 
> > > > > And looking at https://eel.is/c++draft/expr.const#init-4.1 I see that it's
> > > > > even considered constant-initialized.  And that the first note there lies
> > > > > about is_constant_evaluated; I'll email core about that.
> > > > > 
> > > > > So I think what we want is "usable in constant expressions", i.e.
> > > > > decl_constant_var_p.
> > > > 
> > > > I don't really understand.  Even if I keep those two out_of_consteval
> > > > calls, replacing DECL_INITIALIZED_BY_C_E_P with decl_constant_var_p
> > > > doesn't work for e.g.
> > > > 
> > > >     int i = (^^int, 42);
> > > > 
> > > > at namespace scope, where i is DECL_INITIALIZED_BY_C_E_P but not
> > > > decl_constant_var_p.  We can't check the unfolded value here.
> > > 
> > > Ah, right, we need "usable in constant expressions or has constant
> > > initialization" (https://eel.is/c++draft/expr.const#defns-1.6).
> > > 
> > > So decl_constant_var_p or static && BY_C_E_P.
> > 
> > OK, that does look like the right condition, changed.
> > 
> > ...except it still doesn't avoid the need to possibly call
> > out_of_consteval the second time too for the consteval example:
> > 
> >    consteval auto fn () { return ^^int; }
> >    void g() { auto r = fn (); }
> > 
> > `r` isn't decl_constant_var_p and not static so no MCE.  But
> > the unfolded value is a CALL_EXPR so we wouldn't detect the
> > "not associated with a constexpr var" problem.  I suppose I could
> > detect it later in cp_fold_* where we'll see this DECL_EXPR:
> > 
> >    r = REFLECT_EXPR<int>;
> > 
> > but that would mean duplicating the diagnostic and I didn't want to
> > do that.  And another twist is that if `g` is constexpr, we don't
> > eval the fn() call here, and only detect the problem in
> > cp_fold_immediate_r ("If we called a consteval function...") which
> > gets
> > 
> >    r = fn();
> 
> Maybe we want to delay this diagnostic until cp_fold_immediate (rather than
> duplicate it there)?

I'd thought about it but cp_fold_immediate_r only sees the folded
init so we would not give errors for block-scope

  int i = (^^int, 42);

since the initializer then is just 42.

> > diff --git a/gcc/cp/cxx-pretty-print.cc b/gcc/cp/cxx-pretty-print.cc
> > index f5d476a4a99..8f555dd4eab 100644
> > --- a/gcc/cp/cxx-pretty-print.cc
> > +++ b/gcc/cp/cxx-pretty-print.cc
> > @@ -1442,9 +1442,13 @@ cxx_pretty_printer::simple_type_specifier (tree t)
> >         pp_cxx_trait (this, t);
> >         break;
> > -    case META_TYPE:
> > -      pp_cxx_ws_string (this, "std::meta::info");
> > -      break;
> > +    case LANG_TYPE:
> > +      if (REFLECTION_TYPE_P (t))
> > +	{
> > +	  pp_cxx_ws_string (this, "std::meta::info");
> > +	  break;
> > +	}
> > +      gcc_fallthrough ();
> >       default:
> >         c_pretty_printer::simple_type_specifier (t);
> > @@ -1925,6 +1929,13 @@ cxx_pretty_printer::type_id (tree t)
> >     switch (TREE_CODE (t))
> >       {
> > +    case LANG_TYPE:
> > +      if (REFLECTION_TYPE_P (t))
> > +	{
> > +	  pp_cxx_type_specifier_seq (this, t);
> > +	  break;
> > +	}
> > +      goto default_;
> 
> Why not put this just above default: like you did in simple_type_specifier?

Ah sure.  Initially I didn't want to repeat the pp_cxx_type_specifier_seq
call so I just had it fall through to it, but that was too ugly so I just
duplicated it, but didn't move the LANG_TYPE case.  Done now.

> >       case TYPE_DECL:
> >       case UNION_TYPE:
> >       case RECORD_TYPE:
> > @@ -1942,7 +1953,6 @@ cxx_pretty_printer::type_id (tree t)
> >       case NULLPTR_TYPE:
> >       case TEMPLATE_ID_EXPR:
> >       case OFFSET_TYPE:
> > -    case META_TYPE:
> >         pp_cxx_type_specifier_seq (this, t);
> >         if (TYPE_PTRMEM_P (t))
> >   	abstract_declarator (t);
> > @@ -1976,6 +1986,7 @@ cxx_pretty_printer::type_id (tree t)
> >         break;
> >       default:
> > +    default_:
> >         c_pretty_printer::type_id (t);
> >         break;
> >       }
> > diff --git a/gcc/cp/decl.cc b/gcc/cp/decl.cc
> > index 55badc9eed9..6d08e25245d 100644
> > --- a/gcc/cp/decl.cc
> > +++ b/gcc/cp/decl.cc
> > @@ -1008,6 +1008,12 @@ wrapup_namespace_globals ()
> >       {
> >         for (tree decl : *statics)
> >   	{
> > +	  /* Do this here rather than e.g. in make_rtl_for_nonlocal_decl
> > +	     because that would be too early; we need to keep a null
> > +	     reflection as a REFLECT_EXPR for comparisons etc.  */
> > +	  if (flag_reflection && DECL_INITIAL (decl))
> > +	    rewrite_null_reflection (DECL_INITIAL (decl));
> 
> Don't you need to be able to compare a zero to a null reflection to handle
> zero-initialization anyway?

I don't believe so because build_zero_init_1 has:

  else if (REFLECTION_TYPE_P (type))
    /* [dcl.init.general]: "if T is std::meta::info, the object is initialized
       to a null reflection value".  */
    init = get_null_reflection ();

I've added null7.C to exercise this a bit more and found no issues.

Re-ran only dg.exp since there were no significant changes.

-- >8 --
<https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2026/p4101r1.html>
changes the consteval-only type model to a consteval-only values model.
A null reflection is no longer consteval-only.

So, for instance, this is now OK:

  std::meta::info i;
  auto foo (std::meta::info i) { return i; } // not consteval

but

  std::meta::info type = ^^int;

is still ill-formed, provided it's not in an immediate function context

I suppose for users the change is not that dramatic, but our
implementation had to change quite a bit.

For one thing, we now need to be able to handle null reflections in the
ME.  This is done by removing META_TYPE and using LANG_TYPE instead.
In dwarf2out, LANG_TYPE already maps to DW_TAG_unspecified_type
so we should get "decltype(^^int)" in debug info.

To be able to detect invalid code like the bare:

  ^^int;

in cp_fold_*, and to allow code like:

  int i = (^^int, 42); // namespace scope

convert_to_void no longer throws away discarded-value expressions.  But
the modules and analyzer code sort of depended on these being
discarded.  Obviously, first we have to check for consteval-only values
and only then can we discard them.  But we need to do the discarding
before maybe_save_constexpr_fundef; otherwise we'd have to add a separate
walk for the copy.  I moved all of this into cp_fold_r.

As mentioned above, a null reflection can get into the ME, but REFLECT_EXPR
still can't.  rewrite_null_reflection rewrites REFLECT_EXPRs representing
a null reflection to zero.

Taking the address of a consteval function is no longer prohibited outside
an immediate function context, provided the result initializes a constexpr
variable.  I didn't want to inflate the patch with testsuite changes for all
the consteval tests that will need to be updated once this is resolved.

	PR c++/125820

gcc/cp/ChangeLog:

	* constexpr.cc (cxx_eval_outermost_constant_expr): Delete
	consteval-only smuggling detection code.
	(potential_constant_expression_1) <case CONVERT_EXPR>: Recurse
	with want_rval=false for discarded-value expressions.
	* constraint.cc (satisfy_atom): Check require_constant_expression
	before calling cxx_constant_value.  Set result to error_mark_node
	otherwise.
	* cp-gimplify.cc (cp_gimplify_expr) <case REFLECT_EXPR>: New,
	call rewrite_null_reflection on it.
	<case CALL_EXPR>: Don't call consteval_only_p.
	(wipe_consteval_only_r): Remove.
	(cp_fold_immediate_r): Don't detect invalid uses of consteval-only
	types here.
	<case IF_STMT>: Don't call wipe_consteval_only_r.
	(cp_fold_r) <case RETURN_EXPR, EXPR_STMT, STATEMENT_LIST>: New,
	detect invalid uses of consteval-only values here.
	<case CONVERT_EXPR>: New, remove discarded-value expressions.
	(cp_genericize_r) <case BIND_EXPR>: Call rewrite_null_reflection.
	Don't walk BLOCK_VARS here.
	* cp-objcp-common.cc (cp_common_init_ts): Remove the META_TYPE
	marking.
	* cp-tree.def (META_TYPE): Remove.
	* cp-tree.h (REFLECTION_TYPE_P): Adjust to check if TYPE is
	meta_info_type_node.
	(rewrite_null_reflection): Declare.
	(consteval_only_p): No longer pure.
	(check_consteval_only_fn): Remove.
	* cvt.cc (convert_to_void): Don't call check_out_of_consteval_use
	here.  Don't remove discarded-value expressions here.
	* cxx-pretty-print.cc (cxx_pretty_printer::simple_type_specifier)
	<case META_TYPE>: Remove.
	<case LANG_TYPE>: New.
	(cxx_pretty_printer::type_id) <case META_TYPE>: Remove.
	<case LANG_TYPE>: New.
	* decl.cc (wrapup_namespace_globals): Call
	rewrite_null_reflection on each element of statics.
	(cp_finish_decl): Don't call check_out_of_consteval_use.  Don't
	check consteval_only_p.
	(grokfndecl): Don't call check_consteval_only_fn.
	* decl2.cc (maybe_make_one_only): Don't return early for
	consteval_only_p.
	(mark_needed): Likewise.
	(prune_vars_needing_no_initialization): Don't prune
	consteval_only_p variables.
	(c_parse_final_cleanups): Call rewrite_null_reflection.
	* error.cc (dump_type) <case LANG_TYPE>: Handle
	REFLECTION_TYPE_P.
	<case META_TYPE>: Remove.
	(dump_type_prefix) <case META_TYPE>: Remove.
	(dump_type_suffix) <case META_TYPE>: Remove.
	(dump_expr): Use REFLECTION_TYPE_P.
	* init.cc (perform_member_init): Don't call
	check_out_of_consteval_use.
	* mangle.cc (write_type) <case META_TYPE>: Remove.
	<case LANG_TYPE>: Handle REFLECTION_TYPE_P.
	* module.cc (trees_out::type_node) <case META_TYPE>: Remove.
	<case LANG_TYPE>: New.
	(trees_in::tree_node) <case META_TYPE>: Remove.
	<case LANG_TYPE>: New.
	* name-lookup.cc (name_lookup::adl_type): Move the
	REFLECTION_TYPE_P handling to case LANG_TYPE.
	* pt.cc (tsubst) <case META_TYPE>: Remove.
	(unify) <case META_TYPE>: Remove.
	<case LANG_TYPE>: New.
	(instantiate_body): Don't call check_consteval_only_fn.
	* reflect.cc (init_reflection): Use LANG_TYPE instead of
	META_TYPE for meta_info_type_node.
	(null_reflection_p): Also accept the rewritten form.
	(rewrite_null_reflection): New.
	(consteval_only_p): Rewrite for P4101.
	(struct consteval_only_p_walker): Remove.
	(consteval_only_value_p): New.
	(consteval_only_p_walker::walk): Remove.
	(check_out_of_consteval_use_r): Walk INIT_EXPR.  Handle EXPR_STMT
	specially.
	(check_out_of_consteval_use): Remove the special VAR_P handling.
	(check_consteval_only_fn): Remove.
	* search.cc (check_final_overrider): Don't check
	consteval_only_p.
	* semantics.cc (cp_build_bit_cast): Likewise.
	* tree.cc (type_has_unique_obj_representations) <case LANG_TYPE>:
	New, handle REFLECTION_TYPE_P.
	* typeck.cc (cp_build_binary_op): Use REFLECTION_TYPE_P.
	(check_return_expr): Don't call check_out_of_consteval_use.
	* typeck2.cc (store_init_value): Check that an immediate object
	is associated with a constexpr variable.

gcc/ChangeLog:

	* expr.cc (count_type_elements) <case LANG_TYPE>: Return 1.
	* tree.cc (wide_int_to_tree_1) <case LANG_TYPE>: New.
	(cache_integer_cst) <case LANG_TYPE>: New.

libstdc++-v3/ChangeLog:

	* include/std/meta (std::meta::exception::what): Change it to
	constexpr.

gcc/testsuite/ChangeLog:

	* g++.dg/cpp2a/concepts-uneval3.C: Add a dg-error.
	* g++.dg/reflect/bit_cast.C: Remove dg-error.  Extend.
	* g++.dg/reflect/crash12.C: Remove dg-error.
	* g++.dg/reflect/crash18.C: Likewise.
	* g++.dg/reflect/diag3.C: Adjust expected output.
	* g++.dg/reflect/expr11.C: Likewise.
	* g++.dg/reflect/expr12.C: Likewise.
	* g++.dg/reflect/init10.C: Likewise.
	* g++.dg/reflect/init11.C: Likewise.
	* g++.dg/reflect/init12.C: Remove dg-error.
	* g++.dg/reflect/init16.C: Adjust expected output.
	* g++.dg/reflect/init19.C: Remove dg-error.
	* g++.dg/reflect/init4.C: Add an XFAIL.
	* g++.dg/reflect/init5.C: Remove dg-error.
	* g++.dg/reflect/init6.C: Adjust expected output.
	* g++.dg/reflect/init7.C: Likewise.
	* g++.dg/reflect/init9.C: Likewise.
	* g++.dg/reflect/override1.C: Likewise.
	* g++.dg/reflect/p2996-15.C: Remove dg-error.
	* g++.dg/reflect/parm3.C: Adjust expected output.
	* g++.dg/reflect/parm4.C: Likewise.
	* g++.dg/reflect/pr124012.C: Remove dg-error.
	* g++.dg/reflect/type12.C: Likewise.
	* g++.dg/reflect/type2.C: Adjust expected output.
	* g++.dg/reflect/bit_cast2.C: New test.
	* g++.dg/reflect/expr19.C: New test.
	* g++.dg/reflect/null7.C: New test.
	* g++.dg/reflect/type_trait19.C: New test.
	* g++.dg/reflect/typeinfo1.C: New test.
	* g++.dg/reflect/value1.C: New test.
	* g++.dg/reflect/value2.C: New test.
	* g++.dg/reflect/value3.C: New test.
	* g++.dg/reflect/value4.C: New test.
	* g++.dg/reflect/value5.C: New test.
	* g++.dg/reflect/value6.C: New test.
	* g++.dg/reflect/value7.C: New test.
	* g++.dg/reflect/vector2.C: New test.
---
 gcc/cp/constexpr.cc                           |  37 +-
 gcc/cp/constraint.cc                          |   5 +-
 gcc/cp/cp-gimplify.cc                         | 111 ++--
 gcc/cp/cp-objcp-common.cc                     |   1 -
 gcc/cp/cp-tree.def                            |   3 -
 gcc/cp/cp-tree.h                              |   7 +-
 gcc/cp/cvt.cc                                 |  10 -
 gcc/cp/cxx-pretty-print.cc                    |  19 +-
 gcc/cp/decl.cc                                |  23 +-
 gcc/cp/decl2.cc                               |  20 +-
 gcc/cp/error.cc                               |  14 +-
 gcc/cp/init.cc                                |   3 -
 gcc/cp/mangle.cc                              |  11 +-
 gcc/cp/module.cc                              |   8 +-
 gcc/cp/name-lookup.cc                         |  14 +-
 gcc/cp/pt.cc                                  |   5 +-
 gcc/cp/reflect.cc                             | 283 +++++----
 gcc/cp/search.cc                              |  10 +-
 gcc/cp/semantics.cc                           |   6 -
 gcc/cp/tree.cc                                |  10 +
 gcc/cp/typeck.cc                              |   8 +-
 gcc/cp/typeck2.cc                             |  36 ++
 gcc/expr.cc                                   |   3 +-
 gcc/testsuite/g++.dg/cpp2a/concepts-uneval3.C |   2 +-
 gcc/testsuite/g++.dg/reflect/bit_cast.C       |  20 +-
 gcc/testsuite/g++.dg/reflect/bit_cast2.C      |  12 +
 gcc/testsuite/g++.dg/reflect/crash12.C        |   2 +-
 gcc/testsuite/g++.dg/reflect/crash18.C        |   2 +-
 gcc/testsuite/g++.dg/reflect/diag3.C          |  18 +-
 gcc/testsuite/g++.dg/reflect/expr11.C         |  20 +-
 gcc/testsuite/g++.dg/reflect/expr12.C         |  14 +-
 gcc/testsuite/g++.dg/reflect/expr19.C         |   8 +
 gcc/testsuite/g++.dg/reflect/init10.C         |   8 +-
 gcc/testsuite/g++.dg/reflect/init11.C         |   5 +-
 gcc/testsuite/g++.dg/reflect/init12.C         |   2 +-
 gcc/testsuite/g++.dg/reflect/init16.C         |   4 +-
 gcc/testsuite/g++.dg/reflect/init19.C         |   2 +-
 gcc/testsuite/g++.dg/reflect/init4.C          |   2 +-
 gcc/testsuite/g++.dg/reflect/init5.C          |   4 +-
 gcc/testsuite/g++.dg/reflect/init6.C          |  12 +-
 gcc/testsuite/g++.dg/reflect/init7.C          |  22 +-
 gcc/testsuite/g++.dg/reflect/init9.C          |  22 +-
 gcc/testsuite/g++.dg/reflect/null7.C          |  31 +
 gcc/testsuite/g++.dg/reflect/override1.C      |   8 +-
 gcc/testsuite/g++.dg/reflect/p2996-15.C       |   4 +-
 gcc/testsuite/g++.dg/reflect/parm3.C          |  12 +-
 gcc/testsuite/g++.dg/reflect/parm4.C          |  22 +-
 gcc/testsuite/g++.dg/reflect/pr124012.C       |   8 +-
 gcc/testsuite/g++.dg/reflect/type12.C         |   2 +-
 gcc/testsuite/g++.dg/reflect/type2.C          |  14 +-
 gcc/testsuite/g++.dg/reflect/type_trait19.C   |  15 +
 gcc/testsuite/g++.dg/reflect/typeinfo1.C      |   7 +
 gcc/testsuite/g++.dg/reflect/value1.C         | 543 ++++++++++++++++++
 gcc/testsuite/g++.dg/reflect/value2.C         |  32 ++
 gcc/testsuite/g++.dg/reflect/value3.C         |  63 ++
 gcc/testsuite/g++.dg/reflect/value4.C         |  24 +
 gcc/testsuite/g++.dg/reflect/value5.C         |  40 ++
 gcc/testsuite/g++.dg/reflect/value6.C         |  41 ++
 gcc/testsuite/g++.dg/reflect/value7.C         |  52 ++
 gcc/testsuite/g++.dg/reflect/vector2.C        |  26 +
 gcc/tree.cc                                   |   4 +
 libstdc++-v3/include/std/meta                 |   2 +-
 62 files changed, 1337 insertions(+), 441 deletions(-)
 create mode 100644 gcc/testsuite/g++.dg/reflect/bit_cast2.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/expr19.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/null7.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/type_trait19.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/typeinfo1.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/value1.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/value2.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/value3.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/value4.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/value5.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/value6.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/value7.C
 create mode 100644 gcc/testsuite/g++.dg/reflect/vector2.C

base-commit: 0acb69bf9bb0db14b7e230fb5370257c530fb6e8

Message ID	aphVek9XbiEDoRCr@redhat.com
State	New
Headers	show Return-Path: <gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org> X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@legolas.ozlabs.org Authentication-Results: legolas.ozlabs.org; dkim=pass (1024-bit key; unprotected) header.d=redhat.com header.i=@redhat.com header.a=rsa-sha256 header.s=mimecast20190719 header.b=TMlM8Inf; dkim-atps=neutral Authentication-Results: legolas.ozlabs.org; spf=pass (sender SPF authorized) smtp.mailfrom=gcc.gnu.org (client-ip=38.145.34.32; helo=vm01.sourceware.org; envelope-from=gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org; receiver=patchwork.ozlabs.org) Received: from vm01.sourceware.org (vm01.sourceware.org [38.145.34.32]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange x25519 server-signature ECDSA (secp384r1) server-digest SHA384) (No client certificate requested) by legolas.ozlabs.org (Postfix) with ESMTPS id 4hZppR2hq5z1xvf for <incoming@patchwork.ozlabs.org>; Thu, 03 Sep 2026 02:58:39 +1000 (AEST) Received: from vm01.sourceware.org (localhost [IPv6:::1]) by sourceware.org (Postfix) with ESMTP id 69AE84BAE7E6 for <incoming@patchwork.ozlabs.org>; Wed, 2 Sep 2026 16:58:37 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 69AE84BAE7E6 Authentication-Results: sourceware.org; dkim=pass (1024-bit key, unprotected) header.d=redhat.com header.i=@redhat.com header.a=rsa-sha256 header.s=mimecast20190719 header.b=TMlM8Inf X-Original-To: gcc-patches@gcc.gnu.org Delivered-To: gcc-patches@gcc.gnu.org Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by sourceware.org (Postfix) with ESMTP id 393A74B99F5A for <gcc-patches@gcc.gnu.org>; Wed, 2 Sep 2026 16:57:45 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 393A74B99F5A Authentication-Results: sourceware.org; dmarc=pass (p=quarantine dis=none) header.from=redhat.com Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=redhat.com ARC-Filter: OpenARC Filter v1.0.0 sourceware.org 393A74B99F5A Authentication-Results: sourceware.org; arc=none smtp.remote-ip=170.10.129.124 ARC-Seal: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1788368265; cv=none; b=k3qLk0fIFEZNbazG1N9dG9c+3lc5RK12ZMB7AgVA6w3WeppZO3DZ43LhPNsFfHSPhU2u2YeYiR/o2/ACqBwLozLN7VRUJzUBfTk/CkWdHTowjCwfIf0rfdIMPyOi9tsASEwm+bUnkMaUJa57PZJO9as5OtThadMwp5Ghi+vpbOk= ARC-Message-Signature: i=1; a=rsa-sha256; d=sourceware.org; s=key; t=1788368265; c=relaxed/simple; bh=B8F945Dlk6wC88BGiSHzUK6mfixipDM86DHBCHy/GRA=; h=DKIM-Signature:Date:From:To:Subject:Message-ID:MIME-Version; b=f6VH2wOEF73N9CQ4lv1v44ZqlVA7okvkHebn/T9CLvOnJP/bMLEwtekAK7xaW9Y6VAVz72zTuBIIcrsnW/oKk9vF1fuPRfj2GpQk5p4qMBODUIznhGXIFt5rrvqu5Lt8sAm1suR1Pcg5hTRVsVeh5NniDnzM0UsBPYiG/3o6a0Y= ARC-Authentication-Results: i=1; sourceware.org; dkim=pass (1024-bit key, unprotected) header.d=redhat.com header.i=@redhat.com header.a=rsa-sha256 header.s=mimecast20190719 header.b=TMlM8Inf DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 393A74B99F5A DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1788368264; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: in-reply-to:in-reply-to:references:references; bh=5i3CWpF9i2vP422VqbJbAlcpxp3WLaXTFqIWMPIvkGI=; b=TMlM8Infc9Wryr1tECbZxQPuvCdGXrUcf1ufoG4XGC8fqZ/amiUEGwNazxHJYgl1NKN0xF OL25fUsB21w/oijKRT6Z6iV6+U51q1dSTww3WAWRiQtOhq+W+5FjgaZi2EIy+LJ9GCbQfk woCKDZnD+6TerSCIU4Z6Th1E8nbrLlA= Received: from mail-qt1-f199.google.com (mail-qt1-f199.google.com [209.85.160.199]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-219-sbWXtIIdN7aFOQZT5SFzmw-1; Wed, 02 Sep 2026 12:57:42 -0400 X-MC-Unique: sbWXtIIdN7aFOQZT5SFzmw-1 X-Mimecast-MFC-AGG-ID: sbWXtIIdN7aFOQZT5SFzmw_1788368262 Received: by mail-qt1-f199.google.com with SMTP id d75a77b69052e-5283df62d68so22146711cf.0 for <gcc-patches@gcc.gnu.org>; Wed, 02 Sep 2026 09:57:42 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20251104; t=1788368261; x=1788973061; h=user-agent:in-reply-to:content-disposition:content-type :mime-version:references:message-id:subject:cc:to:from:date:x-gm-gg :x-gm-message-state:from:to:cc:subject:date:message-id:reply-to :content-type; bh=5i3CWpF9i2vP422VqbJbAlcpxp3WLaXTFqIWMPIvkGI=; b=BkDY3ZtjP/SO5eTI3LfnakGrtd1XNiKf+0FJaw09IAV8cIRkDhwYGkug6jR28/EZVZ kgCuD597RsB/2Muiw7KP/aYf2z2YVY3YGGYD/RoL8zQhtPwnYV5A9Iq0tg/IXQX1n3Ae 43f8wSk6u1GqAdHBPBGkUG/zmp5CPO2z6WeBcN0AZiHQBuWXvHIcE/l838UT5qQPKaDL f5f9cupwraNMYV/MhmpUrKDHZlEVOmLXy7OyE2Mh0kiTFL7An21REtFBH/fG4Qp7H8oe xaq/VVXueG/SDvWNhWv2QIgJ/CEZBJnQjQ3ygCPj1U5n8SzjPNWHS/YCUTAVU5Byd9Gp jV6A== X-Gm-Message-State: AFuF++mP11YuS57Yqs5xWPKIBzNWhga75LoTUydU31oX2Y2yoNjLZjdD XEBE79aN3gCVo5LJtVADG3EGcf67RRx8KOv7QbVfXNPT5aQkwi3KSrCJy0KJNEUOtMOUXA4Kbt/ Y4EXSTkZ75kiNgF7KHFgIYOosmzug0Y8iAlUt2uZ+1IrshqHej5MbO5RkMEONarIKwns= X-Gm-Gg: AR+sD113jCeXaxUZE8UPS7Lez8EY/jSVykvjJ4NzDwRb3cCKlCS+jLH4+CRxsUxe1rL Oy3RnQbFZG25089jakPwEPrRURzP7Gb296RWBpQaG8Dtw9+sKtk4uvYoyxEGcVGFbaZZ8lE6jbJ 3uHyNS2XaBCPzmkInIjUYj89zkMl5LthEyhiDCgTuqKQrU1wZeeRaj6P6AXlsF0W3zw6anST4BQ JeapSagnGYzYQRuZP96NoOCvYUNUTeM8wKFUV2n7lug+uCktpf4PMCgBO3sANDTvFm3cQpWVVtr E33RRfkPSY4L/Xzi85Xr2l5hH6tk3GWN7oj36b2MOUN+1DHidXSRYY5sxIiqcWjnMjVEhrN+3ay 6Kw== X-Received: by 2002:ac8:7dc2:0:b0:530:4310:5397 with SMTP id d75a77b69052e-53043105654mr6567181cf.47.1788368255337; Wed, 02 Sep 2026 09:57:35 -0700 (PDT) X-Received: by 2002:ac8:7dc2:0:b0:530:4310:5397 with SMTP id d75a77b69052e-53043105654mr6565341cf.47.1788368253055; Wed, 02 Sep 2026 09:57:33 -0700 (PDT) Received: from redhat.com ([2603:7002:1f00:31d2:1607:6e59:bac6:6d08]) by smtp.gmail.com with ESMTPSA id 6a1803df08f44-90e9eec7608sm22096676d6.26.2026.09.02.09.57.31 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 02 Sep 2026 09:57:32 -0700 (PDT) Date: Wed, 2 Sep 2026 12:57:30 -0400 From: Marek Polacek <polacek@redhat.com> To: Jason Merrill <jason@redhat.com> Cc: GCC Patches <gcc-patches@gcc.gnu.org> Subject: [PATCH v4] c++/reflection: P4101, Consteval-only values [PR125820] Message-ID: <aphVek9XbiEDoRCr@redhat.com> References: <20260823220801.1575358-1-polacek@redhat.com> <9b620bbc-21a7-41d9-94de-f2ff88633ef1@redhat.com> <ao3JVhFwSIM4Z4TI@redhat.com> <9d798dfc-5b6c-4e2c-a566-0967763e0d99@redhat.com> <apCh9S2H5OSTi6Rs@redhat.com> <851915fe-14c2-4462-93b2-227d51f3bcdf@redhat.com> <apdkKMN7YKVkIlu2@redhat.com> <0630a587-99c2-413c-87ad-d2ceaa0dc82c@redhat.com> MIME-Version: 1.0 In-Reply-To: <0630a587-99c2-413c-87ad-d2ceaa0dc82c@redhat.com> User-Agent: Mutt/2.4.0 (2026-06-19) X-Mimecast-Spam-Score: 0 X-Mimecast-MFC-PROC-ID: wIUqy4mx-jFQ_XsBmsp4wUAW71-tQb2PhlBAImqFZRU_1788368262 X-Mimecast-Originator: redhat.com Content-Type: text/plain; charset=us-ascii Content-Disposition: inline X-BeenThere: gcc-patches@gcc.gnu.org X-Mailman-Version: 2.1.30 Precedence: list List-Id: Gcc-patches mailing list <gcc-patches.gcc.gnu.org> List-Unsubscribe: <https://gcc.gnu.org/mailman/options/gcc-patches>, <mailto:gcc-patches-request@gcc.gnu.org?subject=unsubscribe> List-Archive: <https://gcc.gnu.org/pipermail/gcc-patches/> List-Post: <mailto:gcc-patches@gcc.gnu.org> List-Help: <mailto:gcc-patches-request@gcc.gnu.org?subject=help> List-Subscribe: <https://gcc.gnu.org/mailman/listinfo/gcc-patches>, <mailto:gcc-patches-request@gcc.gnu.org?subject=subscribe> Errors-To: gcc-patches-bounces~incoming=patchwork.ozlabs.org@gcc.gnu.org
Series	[v4] c++/reflection: P4101, Consteval-only values [PR125820] \| expand [v4] c++/reflection: P4101, Consteval-only values [PR125820]

[v4] c++/reflection: P4101, Consteval-only values [PR125820]

Commit Message

Comments

Patch