c++: ICE with erroneous template redeclaration [PR106311]

Message ID 20220715152920.693196-1-polacek@redhat.com
State New, archived
Headers
Series c++: ICE with erroneous template redeclaration [PR106311] |

Commit Message

Li, Pan2 via Gcc-patches July 15, 2022, 3:29 p.m. UTC
  Here we ICE trying to get DECL_SOURCE_LOCATION of the parm that happens
to be error_mark_node in this ill-formed test.  I kept running into this
while reducing code, so it'd be good to have it fixed.

Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk?

	PR c++/106311

gcc/cp/ChangeLog:

	* pt.cc (redeclare_class_template): Check DECL_P before accessing
	DECL_SOURCE_LOCATION.

gcc/testsuite/ChangeLog:

	* g++.dg/template/redecl5.C: New test.
---
 gcc/cp/pt.cc                            | 3 ++-
 gcc/testsuite/g++.dg/template/redecl5.C | 5 +++++
 2 files changed, 7 insertions(+), 1 deletion(-)
 create mode 100644 gcc/testsuite/g++.dg/template/redecl5.C


base-commit: 23dd41c480fa9f06c33c1e6090bbae53869f85af
  

Comments

Li, Pan2 via Gcc-patches July 22, 2022, 12:35 p.m. UTC | #1
Ping.

On Fri, Jul 15, 2022 at 11:29:20AM -0400, Marek Polacek via Gcc-patches wrote:
> Here we ICE trying to get DECL_SOURCE_LOCATION of the parm that happens
> to be error_mark_node in this ill-formed test.  I kept running into this
> while reducing code, so it'd be good to have it fixed.
> 
> Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk?
> 
> 	PR c++/106311
> 
> gcc/cp/ChangeLog:
> 
> 	* pt.cc (redeclare_class_template): Check DECL_P before accessing
> 	DECL_SOURCE_LOCATION.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.dg/template/redecl5.C: New test.
> ---
>  gcc/cp/pt.cc                            | 3 ++-
>  gcc/testsuite/g++.dg/template/redecl5.C | 5 +++++
>  2 files changed, 7 insertions(+), 1 deletion(-)
>  create mode 100644 gcc/testsuite/g++.dg/template/redecl5.C
> 
> diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
> index 718dfa5bfa8..0a294e91a79 100644
> --- a/gcc/cp/pt.cc
> +++ b/gcc/cp/pt.cc
> @@ -6377,7 +6377,8 @@ redeclare_class_template (tree type, tree parms, tree cons)
>  	{
>  	  auto_diagnostic_group d;
>  	  error ("template parameter %q+#D", tmpl_parm);
> -	  inform (DECL_SOURCE_LOCATION (parm), "redeclared here as %q#D", parm);
> +	  inform (DECL_P (parm) ? DECL_SOURCE_LOCATION (parm) : input_location,
> +		  "redeclared here as %q#D", parm);
>  	  return false;
>  	}
>  
> diff --git a/gcc/testsuite/g++.dg/template/redecl5.C b/gcc/testsuite/g++.dg/template/redecl5.C
> new file mode 100644
> index 00000000000..fb2d698e6bc
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/template/redecl5.C
> @@ -0,0 +1,5 @@
> +// PR c++/106311
> +// { dg-do compile }
> +
> +template <typename, long> struct array; // { dg-error "template parameter" }
> +template <typename, size_t X> struct array { }; // { dg-error "declared" }
> 
> base-commit: 23dd41c480fa9f06c33c1e6090bbae53869f85af
> -- 
> 2.36.1
> 

Marek
  
Li, Pan2 via Gcc-patches July 22, 2022, 9:21 p.m. UTC | #2
On 7/15/22 11:29, Marek Polacek wrote:
> Here we ICE trying to get DECL_SOURCE_LOCATION of the parm that happens
> to be error_mark_node in this ill-formed test.  I kept running into this
> while reducing code, so it'd be good to have it fixed.
> 
> Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk?
> 
> 	PR c++/106311
> 
> gcc/cp/ChangeLog:
> 
> 	* pt.cc (redeclare_class_template): Check DECL_P before accessing
> 	DECL_SOURCE_LOCATION.
> 
> gcc/testsuite/ChangeLog:
> 
> 	* g++.dg/template/redecl5.C: New test.
> ---
>   gcc/cp/pt.cc                            | 3 ++-
>   gcc/testsuite/g++.dg/template/redecl5.C | 5 +++++
>   2 files changed, 7 insertions(+), 1 deletion(-)
>   create mode 100644 gcc/testsuite/g++.dg/template/redecl5.C
> 
> diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
> index 718dfa5bfa8..0a294e91a79 100644
> --- a/gcc/cp/pt.cc
> +++ b/gcc/cp/pt.cc
> @@ -6377,7 +6377,8 @@ redeclare_class_template (tree type, tree parms, tree cons)
>   	{
>   	  auto_diagnostic_group d;
>   	  error ("template parameter %q+#D", tmpl_parm);
> -	  inform (DECL_SOURCE_LOCATION (parm), "redeclared here as %q#D", parm);
> +	  inform (DECL_P (parm) ? DECL_SOURCE_LOCATION (parm) : input_location,
> +		  "redeclared here as %q#D", parm);

If we're checking DECL_P, probably we also should avoid passing it to 
%q#D if it's false?

>   	  return false;
>   	}
>   
> diff --git a/gcc/testsuite/g++.dg/template/redecl5.C b/gcc/testsuite/g++.dg/template/redecl5.C
> new file mode 100644
> index 00000000000..fb2d698e6bc
> --- /dev/null
> +++ b/gcc/testsuite/g++.dg/template/redecl5.C
> @@ -0,0 +1,5 @@
> +// PR c++/106311
> +// { dg-do compile }
> +
> +template <typename, long> struct array; // { dg-error "template parameter" }
> +template <typename, size_t X> struct array { }; // { dg-error "declared" }
> 
> base-commit: 23dd41c480fa9f06c33c1e6090bbae53869f85af
  

Patch

diff --git a/gcc/cp/pt.cc b/gcc/cp/pt.cc
index 718dfa5bfa8..0a294e91a79 100644
--- a/gcc/cp/pt.cc
+++ b/gcc/cp/pt.cc
@@ -6377,7 +6377,8 @@  redeclare_class_template (tree type, tree parms, tree cons)
 	{
 	  auto_diagnostic_group d;
 	  error ("template parameter %q+#D", tmpl_parm);
-	  inform (DECL_SOURCE_LOCATION (parm), "redeclared here as %q#D", parm);
+	  inform (DECL_P (parm) ? DECL_SOURCE_LOCATION (parm) : input_location,
+		  "redeclared here as %q#D", parm);
 	  return false;
 	}
 
diff --git a/gcc/testsuite/g++.dg/template/redecl5.C b/gcc/testsuite/g++.dg/template/redecl5.C
new file mode 100644
index 00000000000..fb2d698e6bc
--- /dev/null
+++ b/gcc/testsuite/g++.dg/template/redecl5.C
@@ -0,0 +1,5 @@ 
+// PR c++/106311
+// { dg-do compile }
+
+template <typename, long> struct array; // { dg-error "template parameter" }
+template <typename, size_t X> struct array { }; // { dg-error "declared" }