* [PATCH] c++: Diagnose this specifier on template parameters [PR113929]
@ 2024-02-16 9:21 Jakub Jelinek
2024-02-16 15:15 ` Jason Merrill
0 siblings, 1 reply; 2+ messages in thread
From: Jakub Jelinek @ 2024-02-16 9:21 UTC (permalink / raw)
To: Jason Merrill; +Cc: gcc-patches
Hi!
For template parameters, the optional this specifier is in the grammar
template-parameter-list -> template-parameter -> parameter-declaration,
just [dcl.fct/6] says that it is only valid in parameter-list of certain
functions. So, unlike the case of decl-specifier-seq used in non-terminals
other than parameter-declaration, I think it is better not to fix this
by
cp_parser_decl_specifier_seq (parser,
- flags | CP_PARSER_FLAGS_PARAMETER,
+ flags | (template_parameter_p ? 0
+ : CP_PARSER_FLAGS_PARAMETER),
&decl_specifiers,
&declares_class_or_enum);
which would be pretending it isn't in the grammar, but by diagnosing it
separately, which is what the following patch does.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2024-02-16 Jakub Jelinek <jakub@redhat.com>
PR c++/113929
* parser.cc (cp_parser_parameter_declaration): Diagnose this specifier
on template parameter declaration.
* g++.dg/parse/pr113929.C: New test.
--- gcc/cp/parser.cc.jj 2024-02-15 17:33:11.641453437 +0100
+++ gcc/cp/parser.cc 2024-02-15 17:40:29.592447265 +0100
@@ -25724,8 +25724,15 @@ cp_parser_parameter_declaration (cp_pars
for a C-style variadic function. */
token = cp_lexer_peek_token (parser->lexer);
- bool const xobj_param_p
+ bool xobj_param_p
= decl_spec_seq_has_spec_p (&decl_specifiers, ds_this);
+ if (xobj_param_p && template_parm_p)
+ {
+ error_at (decl_specifiers.locations[ds_this],
+ "%<this%> specifier in template parameter declaration");
+ xobj_param_p = false;
+ decl_specifiers.locations[ds_this] = 0;
+ }
bool diag_xobj_parameter_pack = false;
if (xobj_param_p && (declarator && declarator->parameter_pack_p))
--- gcc/testsuite/g++.dg/parse/pr113929.C.jj 2024-02-15 17:43:18.500129688 +0100
+++ gcc/testsuite/g++.dg/parse/pr113929.C 2024-02-15 17:42:54.564458109 +0100
@@ -0,0 +1,7 @@
+// PR c++/113929
+// { dg-do compile }
+
+template <this int C> // { dg-error "'this' specifier in template parameter declaration" }
+struct S {};
+template <int N, this int C> // { dg-error "'this' specifier in template parameter declaration" }
+struct T {};
Jakub
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] c++: Diagnose this specifier on template parameters [PR113929]
2024-02-16 9:21 [PATCH] c++: Diagnose this specifier on template parameters [PR113929] Jakub Jelinek
@ 2024-02-16 15:15 ` Jason Merrill
0 siblings, 0 replies; 2+ messages in thread
From: Jason Merrill @ 2024-02-16 15:15 UTC (permalink / raw)
To: Jakub Jelinek; +Cc: gcc-patches
On 2/16/24 04:21, Jakub Jelinek wrote:
> Hi!
>
> For template parameters, the optional this specifier is in the grammar
> template-parameter-list -> template-parameter -> parameter-declaration,
> just [dcl.fct/6] says that it is only valid in parameter-list of certain
> functions. So, unlike the case of decl-specifier-seq used in non-terminals
> other than parameter-declaration, I think it is better not to fix this
> by
> cp_parser_decl_specifier_seq (parser,
> - flags | CP_PARSER_FLAGS_PARAMETER,
> + flags | (template_parameter_p ? 0
> + : CP_PARSER_FLAGS_PARAMETER),
> &decl_specifiers,
> &declares_class_or_enum);
> which would be pretending it isn't in the grammar, but by diagnosing it
> separately, which is what the following patch does.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
OK.
> 2024-02-16 Jakub Jelinek <jakub@redhat.com>
>
> PR c++/113929
> * parser.cc (cp_parser_parameter_declaration): Diagnose this specifier
> on template parameter declaration.
>
> * g++.dg/parse/pr113929.C: New test.
>
> --- gcc/cp/parser.cc.jj 2024-02-15 17:33:11.641453437 +0100
> +++ gcc/cp/parser.cc 2024-02-15 17:40:29.592447265 +0100
> @@ -25724,8 +25724,15 @@ cp_parser_parameter_declaration (cp_pars
> for a C-style variadic function. */
> token = cp_lexer_peek_token (parser->lexer);
>
> - bool const xobj_param_p
> + bool xobj_param_p
> = decl_spec_seq_has_spec_p (&decl_specifiers, ds_this);
> + if (xobj_param_p && template_parm_p)
> + {
> + error_at (decl_specifiers.locations[ds_this],
> + "%<this%> specifier in template parameter declaration");
> + xobj_param_p = false;
> + decl_specifiers.locations[ds_this] = 0;
> + }
>
> bool diag_xobj_parameter_pack = false;
> if (xobj_param_p && (declarator && declarator->parameter_pack_p))
> --- gcc/testsuite/g++.dg/parse/pr113929.C.jj 2024-02-15 17:43:18.500129688 +0100
> +++ gcc/testsuite/g++.dg/parse/pr113929.C 2024-02-15 17:42:54.564458109 +0100
> @@ -0,0 +1,7 @@
> +// PR c++/113929
> +// { dg-do compile }
> +
> +template <this int C> // { dg-error "'this' specifier in template parameter declaration" }
> +struct S {};
> +template <int N, this int C> // { dg-error "'this' specifier in template parameter declaration" }
> +struct T {};
>
> Jakub
>
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2024-02-16 15:16 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-02-16 9:21 [PATCH] c++: Diagnose this specifier on template parameters [PR113929] Jakub Jelinek
2024-02-16 15:15 ` Jason Merrill
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).