From: Jason Merrill <jason@redhat.com>
To: Jakub Jelinek <jakub@redhat.com>
Cc: Jonathan Wakely <jwakely@redhat.com>,
gcc-patches@gcc.gnu.org, Marek Polacek <polacek@redhat.com>
Subject: Re: [PATCH] c++, v3: Implement C++23 P2647R1 - Permitting static constexpr variables in constexpr functions
Date: Thu, 17 Nov 2022 09:42:18 -0500 [thread overview]
Message-ID: <740b5e1e-7143-c291-5594-af937867fbc3@redhat.com> (raw)
In-Reply-To: <Y3X7UH00hQtTnQSj@tucnak>
On 11/17/22 04:13, Jakub Jelinek wrote:
> On Wed, Nov 16, 2022 at 03:26:32PM -0500, Jason Merrill wrote:
>> On 11/16/22 09:46, Jakub Jelinek wrote:
>>> On Wed, Nov 16, 2022 at 09:33:27AM -0500, Jason Merrill wrote:
>>>>> and at that point I fear decl_maybe_constant_var_p will not work
>>>>> properly. Shall this hunk be moved somewhere else (cp_finish_decl?)
>>>>> where we can already call it, or do the above in start_decl for
>>>>> cxx_dialect < cxx20 and add a cxx_dialect == cxx20 hunk in cp_finish_decl?
>>>>
>>>> Hmm, I'd expect decl_maybe_constant_var_p to work fine at this point.
>>>
>>> For static constexpr vars sure, but what about
>>> static const where start_decl doesn't know the initializer?
>>> Sure, decl_maybe_constant_var_p will not crash in that case, but
>>> it will return true even if the static const var doesn't have
>>> a constant initializer. Sure, we'd catch that later on when actually
>>> trying to constexpr evaluate the function and hitting there the
>>> spots added for C++23 in potential_constant_expression*/cxx_eval_*,
>>> but it would mean that we don't reject it when nothing calls the functions.
>>>
>>> I meant something like:
>>> constexpr int bar (int x) { if (x) throw 1; return 0; }
>>> constexpr int foo () { static const int a = bar (1); return 0; }
>>> with -std=c++20 IMHO shouldn't be accepted, while in C++23 it should.
>>
>> I'd expect us to reject that in C++20 in potential_constant_expression, but
>> it's a fair point; it is awkward that P2242 wasn't also accepted as a DR.
I pointed out that inconsistency on the reflectors, and it seems that
EWG didn't actually intend P2647 to be a DR, either; it was a CWG error
to propose it as such. But let's go ahead and build on the work you've
done.
>> Moving the check from start_decl to cp_finish_decl makes sense to me.
>
> So like this? I had to outline the check from start_decl to a function
> because it is needed in cp_finish_decl in two different places (the
> processing_template_decl path and at the end and it can't be done before the
> processing_template_decl spot, because the initializer isn't finalized at
> that point for !procesing_template_decl). Also, decl_maybe_constant_var_p
> doesn't do what is needed when !processing_template_decl, but I think
> we want decl_maybe_constant_var_p in templates so that we don't instantiate
> anything.
>
> Lightly tested so far, ok for trunk if it passes full bootstrap/regtest?
>
> 2022-11-17 Jakub Jelinek <jakub@redhat.com>
>
> gcc/c-family/
> * c-cppbuiltin.cc (c_cpp_builtins): Bump __cpp_constexpr
> value from 202207L to 202211L.
> gcc/cp/
> * constexpr.cc (cxx_eval_constant_expression): Implement C++23
> P2647R1 - Permitting static constexpr variables in constexpr functions.
> Allow decl_constant_var_p static or thread_local vars for
> C++20 and later.
> (potential_constant_expression_1): For C++20 or later, allow
> static or thread_local decl_maybe_constant_var_p vars, for
> !processing_template_decl only decl_constant_var_p vars.
> * decl.cc (diagnose_static_in_constexpr): New function.
> (start_decl): Use it for C++17 or earlier.
> (cp_finish_decl): Call it for C++20.
> gcc/testsuite/
> * g++.dg/cpp23/constexpr-nonlit17.C: New test.
> * g++.dg/cpp23/constexpr-nonlit18.C: New test.
> * g++.dg/cpp23/feat-cxx2b.C: Adjust expected __cpp_constexpr
> value.
> * g++.dg/ext/stmtexpr19.C: Don't expect an error for C++20 or later.
>
> --- gcc/c-family/c-cppbuiltin.cc.jj 2022-11-17 09:00:42.106249011 +0100
> +++ gcc/c-family/c-cppbuiltin.cc 2022-11-17 09:01:49.286320527 +0100
> @@ -1074,7 +1074,7 @@ c_cpp_builtins (cpp_reader *pfile)
> /* Set feature test macros for C++23. */
> cpp_define (pfile, "__cpp_size_t_suffix=202011L");
> cpp_define (pfile, "__cpp_if_consteval=202106L");
> - cpp_define (pfile, "__cpp_constexpr=202207L");
> + cpp_define (pfile, "__cpp_constexpr=202211L");
> cpp_define (pfile, "__cpp_multidimensional_subscript=202211L");
> cpp_define (pfile, "__cpp_named_character_escapes=202207L");
> cpp_define (pfile, "__cpp_static_call_operator=202207L");
> --- gcc/cp/constexpr.cc.jj 2022-11-17 08:48:30.530357181 +0100
> +++ gcc/cp/constexpr.cc 2022-11-17 09:56:50.479522863 +0100
> @@ -7098,7 +7098,8 @@ cxx_eval_constant_expression (const cons
> && (TREE_STATIC (r)
> || (CP_DECL_THREAD_LOCAL_P (r) && !DECL_REALLY_EXTERN (r)))
> /* Allow __FUNCTION__ etc. */
> - && !DECL_ARTIFICIAL (r))
> + && !DECL_ARTIFICIAL (r)
> + && (cxx_dialect < cxx20 || !decl_constant_var_p (r)))
I don't think we need to check cxx_dialect here since
diagnose_static_in_constexpr will have already complained.
> {
> if (!ctx->quiet)
> {
> @@ -9588,7 +9589,12 @@ potential_constant_expression_1 (tree t,
> tmp = DECL_EXPR_DECL (t);
> if (VAR_P (tmp) && !DECL_ARTIFICIAL (tmp))
> {
> - if (CP_DECL_THREAD_LOCAL_P (tmp) && !DECL_REALLY_EXTERN (tmp))
> + if (CP_DECL_THREAD_LOCAL_P (tmp)
> + && !DECL_REALLY_EXTERN (tmp)
> + && (cxx_dialect < cxx20
> + || (processing_template_decl
> + ? !decl_maybe_constant_var_p (tmp)
> + : !decl_constant_var_p (tmp))))
Or here.
> {
> if (flags & tf_error)
> constexpr_error (DECL_SOURCE_LOCATION (tmp), fundef_p,
> @@ -9596,7 +9602,11 @@ potential_constant_expression_1 (tree t,
> "%<constexpr%> context", tmp);
> return false;
> }
> - else if (TREE_STATIC (tmp))
> + else if (TREE_STATIC (tmp)
> + && (cxx_dialect < cxx20
> + || (processing_template_decl
> + ? !decl_maybe_constant_var_p (tmp)
> + : !decl_constant_var_p (tmp))))
> {
> if (flags & tf_error)
> constexpr_error (DECL_SOURCE_LOCATION (tmp), fundef_p,
> --- gcc/cp/decl.cc.jj 2022-11-16 14:44:43.692339668 +0100
> +++ gcc/cp/decl.cc 2022-11-17 09:54:30.556482833 +0100
> @@ -5600,6 +5600,41 @@ groktypename (cp_decl_specifier_seq *typ
> return type;
> }
>
> +/* For C++17 and older diagnose static or thread_local decls in constexpr
> + or consteval functions. For C++20 similarly, except if they are
> + usable in constant expressions. */
> +
> +static void
> +diagnose_static_in_constexpr (tree decl)
> +{
> + if (current_function_decl && VAR_P (decl)
> + && DECL_DECLARED_CONSTEXPR_P (current_function_decl)
> + && cxx_dialect < cxx23
> + && (cxx_dialect < cxx20
> + || (processing_template_decl
> + ? !decl_maybe_constant_var_p (decl)
> + : !decl_constant_var_p (decl))))
For (maybe) constant variables let's make this error a pedwarn in C++20
and below.
> + {
> + bool ok = false;
> + if (CP_DECL_THREAD_LOCAL_P (decl) && !DECL_REALLY_EXTERN (decl))
> + error_at (DECL_SOURCE_LOCATION (decl),
> + "%qD defined %<thread_local%> in %qs function only "
> + "available with %<-std=c++2b%> or %<-std=gnu++2b%>", decl,
> + DECL_IMMEDIATE_FUNCTION_P (current_function_decl)
> + ? "consteval" : "constexpr");
> + else if (TREE_STATIC (decl))
> + error_at (DECL_SOURCE_LOCATION (decl),
> + "%qD defined %<static%> in %qs function only available "
> + "with %<-std=c++2b%> or %<-std=gnu++2b%>", decl,
> + DECL_IMMEDIATE_FUNCTION_P (current_function_decl)
> + ? "consteval" : "constexpr");
> + else
> + ok = true;
> + if (!ok)
> + cp_function_chain->invalid_constexpr = true;
> + }
> +}
> +
> /* Process a DECLARATOR for a function-scope or namespace-scope
> variable or function declaration.
> (Function definitions go through start_function; class member
> @@ -5860,28 +5895,8 @@ start_decl (const cp_declarator *declara
> DECL_THIS_STATIC (decl) = 1;
> }
>
> - if (current_function_decl && VAR_P (decl)
> - && DECL_DECLARED_CONSTEXPR_P (current_function_decl)
> - && cxx_dialect < cxx23)
> - {
> - bool ok = false;
> - if (CP_DECL_THREAD_LOCAL_P (decl) && !DECL_REALLY_EXTERN (decl))
> - error_at (DECL_SOURCE_LOCATION (decl),
> - "%qD defined %<thread_local%> in %qs function only "
> - "available with %<-std=c++2b%> or %<-std=gnu++2b%>", decl,
> - DECL_IMMEDIATE_FUNCTION_P (current_function_decl)
> - ? "consteval" : "constexpr");
> - else if (TREE_STATIC (decl))
> - error_at (DECL_SOURCE_LOCATION (decl),
> - "%qD defined %<static%> in %qs function only available "
> - "with %<-std=c++2b%> or %<-std=gnu++2b%>", decl,
> - DECL_IMMEDIATE_FUNCTION_P (current_function_decl)
> - ? "consteval" : "constexpr");
> - else
> - ok = true;
> - if (!ok)
> - cp_function_chain->invalid_constexpr = true;
> - }
> + if (cxx_dialect < cxx20)
> + diagnose_static_in_constexpr (decl);
Can we drop this call (and make the ones in cp_finish_decl unconditional)?
>
> if (!processing_template_decl && VAR_P (decl))
> start_decl_1 (decl, initialized);
> @@ -8424,6 +8439,10 @@ cp_finish_decl (tree decl, tree init, bo
> set_user_assembler_name (decl, asmspec);
> DECL_HARD_REGISTER (decl) = 1;
> }
> +
> + if (cxx_dialect == cxx20)
> + diagnose_static_in_constexpr (decl);
> +
> return;
> }
>
> @@ -8749,6 +8768,9 @@ cp_finish_decl (tree decl, tree init, bo
> && !DECL_HARD_REGISTER (decl))
> targetm.lower_local_decl_alignment (decl);
>
> + if (cxx_dialect == cxx20)
> + diagnose_static_in_constexpr (decl);
> +
> invoke_plugin_callbacks (PLUGIN_FINISH_DECL, decl);
> }
>
> --- gcc/testsuite/g++.dg/cpp23/constexpr-nonlit17.C.jj 2022-11-17 09:00:42.108248984 +0100
> +++ gcc/testsuite/g++.dg/cpp23/constexpr-nonlit17.C 2022-11-17 09:00:42.108248984 +0100
> @@ -0,0 +1,12 @@
> +// P2647R1 - Permitting static constexpr variables in constexpr functions
> +// { dg-do compile { target c++20 } }
> +
> +constexpr char
> +test ()
> +{
> + static const int x = 5;
> + static constexpr char c[] = "Hello World";
> + return *(c + x);
> +}
> +
> +static_assert (test () == ' ');
> --- gcc/testsuite/g++.dg/cpp23/constexpr-nonlit18.C.jj 2022-11-17 09:29:45.776136195 +0100
> +++ gcc/testsuite/g++.dg/cpp23/constexpr-nonlit18.C 2022-11-17 10:04:32.894045579 +0100
> @@ -0,0 +1,49 @@
> +// P2647R1 - Permitting static constexpr variables in constexpr functions
> +// { dg-do compile { target c++14 } }
> +
> +constexpr int
> +f1 (int x)
> +{
> + if (x)
> + throw 1;
> + return 0;
> +}
> +
> +constexpr int
> +f2 ()
> +{
> + static const int a = f1 (1); // { dg-error "'a' defined 'static' in 'constexpr' function only available with" "" { target c++20_down } }
> + return 0;
> +}
> +
> +constexpr int
> +f3 ()
> +{
> + static const int a = 5; // { dg-error "'a' defined 'static' in 'constexpr' function only available with" "" { target c++17_down } }
> + return 0;
> +}
> +
> +constexpr int
> +f4 () // { dg-message "declared here" "" { target c++20_down } }
> +{ // { dg-message "is not usable as a 'constexpr' function because:" "" { target c++23 } .-1 }
> + static const int a = f1 (1); // { dg-error "'a' defined 'static' in 'constexpr' function only available with" "" { target c++20_down } }
> + return 0; // { dg-error "'a' defined 'static' in 'constexpr' context" "" { target c++23 } .-1 }
> +}
> +
> +constexpr int a4 = f4 (); // { dg-error "called in a constant expression" }
> +
> +constexpr int
> +f5 ()
> +{
> + static const int a = f1 (0); // { dg-error "'a' defined 'static' in 'constexpr' function only available with" "" { target c++17_down } }
> + return 0;
> +}
> +
> +constexpr int
> +f6 () // { dg-message "declared here" "" { target c++17_down } }
> +{
> + static const int a = f1 (0); // { dg-error "'a' defined 'static' in 'constexpr' function only available with" "" { target c++17_down } }
> + return 0;
> +}
> +
> +constexpr int a6 = f6 (); // { dg-error "called in a constant expression" "" { target c++17_down } }
> --- gcc/testsuite/g++.dg/cpp23/feat-cxx2b.C.jj 2022-11-17 08:48:30.561356753 +0100
> +++ gcc/testsuite/g++.dg/cpp23/feat-cxx2b.C 2022-11-17 09:00:42.108248984 +0100
> @@ -134,8 +134,8 @@
>
> #ifndef __cpp_constexpr
> # error "__cpp_constexpr"
> -#elif __cpp_constexpr != 202207
> -# error "__cpp_constexpr != 202207"
> +#elif __cpp_constexpr != 202211
> +# error "__cpp_constexpr != 202211"
> #endif
>
> #ifndef __cpp_decltype_auto
> --- gcc/testsuite/g++.dg/ext/stmtexpr19.C.jj 2022-11-17 08:48:02.730741221 +0100
> +++ gcc/testsuite/g++.dg/ext/stmtexpr19.C 2022-11-17 09:00:42.109248970 +0100
> @@ -8,7 +8,7 @@ const test* setup()
> {
> static constexpr test atest =
> {
> - ({ static const int inner = 123; &inner; }) // { dg-error "static" }
> + ({ static const int inner = 123; &inner; }) // { dg-error "static" "" { target c++17_down } }
> };
>
> return &atest;
>
>
> Jakub
>
next prev parent reply other threads:[~2022-11-17 14:42 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-11 17:07 [PATCH] c++: " Jakub Jelinek
2022-11-13 11:45 ` [PATCH] c++, v2: " Jakub Jelinek
2022-11-15 23:36 ` Jason Merrill
2022-11-15 23:50 ` Jakub Jelinek
2022-11-16 0:27 ` Jonathan Wakely
2022-11-16 6:19 ` Jakub Jelinek
2022-11-16 13:20 ` Jason Merrill
2022-11-16 14:08 ` Jakub Jelinek
2022-11-16 14:33 ` Jason Merrill
2022-11-16 14:46 ` Jakub Jelinek
2022-11-16 20:26 ` Jason Merrill
2022-11-17 9:13 ` [PATCH] c++, v3: " Jakub Jelinek
2022-11-17 14:42 ` Jason Merrill [this message]
2022-11-17 18:42 ` Jakub Jelinek
2022-11-17 20:42 ` [PATCH] c++, v4: " Jakub Jelinek
2022-11-18 0:15 ` Marek Polacek
2022-11-18 7:48 ` Jakub Jelinek
2022-11-18 15:03 ` Marek Polacek
2022-11-18 15:14 ` Jakub Jelinek
2022-11-18 16:24 ` Jason Merrill
2022-11-18 16:34 ` Jakub Jelinek
2022-11-18 16:52 ` Jason Merrill
2022-11-18 0:28 ` Jason Merrill
2022-11-18 9:10 ` [PATCH] c++, v5: " Jakub Jelinek
2022-11-16 0:26 ` [PATCH] c++, v2: " Jonathan Wakely
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=740b5e1e-7143-c291-5594-af937867fbc3@redhat.com \
--to=jason@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jakub@redhat.com \
--cc=jwakely@redhat.com \
--cc=polacek@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).