[C++ PATCH] Avoid ice-on-invalid in xref_basetypes (PR c++/34089)

[C++ PATCH] Avoid ice-on-invalid in xref_basetypes (PR c++/34089)

[Jakub Jelinek]

Hi!

xref_basetypes relies on the first argument being aggregate type,
but on invalid testcases like the attached one cp_parser_class_specifier
can contain various other types (in this case LANG_TYPE - a type
of OVERLOAD).
As begin_class_definition has all the needed code to report diagnostics
and have sane error recovery and cp_parser_class_specifier calls it shortly
after this, I think the best fix for this is instead of duplicating
such checks and error recovery just not call xref_basetypes for
non-aggregates at all and just let begin_class_definition do its job.

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

2007-11-19  Jakub Jelinek  <jakub@redhat.com>

	PR c++/34089
	* parser.c (cp_parser_class_specifier): Don't call xref_basetypes
	if type is not aggregate type.

	* g++.dg/template/crash74.C: New test.

--- gcc/cp/parser.c.jj	2007-11-08 01:19:18.000000000 +0100
+++ gcc/cp/parser.c	2007-11-19 17:47:34.000000000 +0100
@@ -14134,7 +14134,7 @@ cp_parser_class_specifier (cp_parser* pa
 
   /* Process the base classes. If they're invalid, skip the 
      entire class body.  */
-  if (!xref_basetypes (type, bases))
+  if (IS_AGGR_TYPE (type) && !xref_basetypes (type, bases))
     {
       /* Consuming the closing brace yields better error messages
          later on.  */
--- gcc/testsuite/g++.dg/template/crash74.C.jj	2007-11-19 17:53:27.000000000 +0100
+++ gcc/testsuite/g++.dg/template/crash74.C	2007-11-19 17:54:38.000000000 +0100
@@ -0,0 +1,6 @@
+// PR c++/34089
+// { dg-do compile }
+// { dg-options "" }
+
+template<typename F> void foo () { }
+template<typename F> struct foo<F> { };	// { dg-error "template class without a name|abstract decl" }

	Jakub

Re: [C++ PATCH] Avoid ice-on-invalid in xref_basetypes (PR c++/34089)

[Mark Mitchell]

Jakub Jelinek wrote:

> 2007-11-19  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR c++/34089
> 	* parser.c (cp_parser_class_specifier): Don't call xref_basetypes
> 	if type is not aggregate type.
> 
> 	* g++.dg/template/crash74.C: New test.

This seems like it would work, but why are we event accepting this code
in cp_parser_class_head?

+template<typename F> void foo () { }
+template<typename F> struct foo<F> { };

Here, we know that foo is not a name for a class template, so we should
be able to detect the error at that point.

Thanks,

-- 
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713

[C++ PATCH] Avoid ice-on-invalid in xref_basetypes (PR c++/34089, take 2)

[Jakub Jelinek]

On Mon, Nov 19, 2007 at 09:01:58PM -0800, Mark Mitchell wrote:
> Jakub Jelinek wrote:
> 
> > 2007-11-19  Jakub Jelinek  <jakub@redhat.com>
> > 
> > 	PR c++/34089
> > 	* parser.c (cp_parser_class_specifier): Don't call xref_basetypes
> > 	if type is not aggregate type.
> > 
> > 	* g++.dg/template/crash74.C: New test.
> 
> This seems like it would work, but why are we event accepting this code
> in cp_parser_class_head?
> 
> +template<typename F> void foo () { }
> +template<typename F> struct foo<F> { };
> 
> Here, we know that foo is not a name for a class template, so we should
> be able to detect the error at that point.

Like this?

2007-11-20  Jakub Jelinek  <jakub@redhat.com>

	PR c++/34089
	* parser.c (cp_parser_class_head): Reject function template ids.

	* g++.dg/template/crash74.C: New test.

--- gcc/cp/parser.c.jj	2007-11-20 11:31:07.000000000 +0100
+++ gcc/cp/parser.c	2007-11-20 14:51:47.000000000 +0100
@@ -14536,8 +14536,18 @@ cp_parser_class_head (cp_parser* parser,
   /* Look up the type.  */
   if (template_id_p)
     {
-      type = TREE_TYPE (id);
-      type = maybe_process_partial_specialization (type);
+      if (TREE_CODE (id) == TEMPLATE_ID_EXPR
+	  && (DECL_FUNCTION_TEMPLATE_P (TREE_OPERAND (id, 0))
+	      || TREE_CODE (TREE_OPERAND (id, 0)) == OVERLOAD))
+	{
+	  error ("function template %qD redeclared as a class template", id);
+	  type = error_mark_node;
+	}
+      else
+	{
+	  type = TREE_TYPE (id);
+	  type = maybe_process_partial_specialization (type);
+	}
       if (nested_name_specifier)
 	pushed_scope = push_scope (nested_name_specifier);
     }
--- gcc/testsuite/g++.dg/template/crash74.C.jj	2007-11-20 14:08:53.000000000 +0100
+++ gcc/testsuite/g++.dg/template/crash74.C	2007-11-20 14:52:25.000000000 +0100
@@ -0,0 +1,6 @@
+// PR c++/34089
+// { dg-do compile }
+// { dg-options "" }
+
+template<typename F> void foo () { }
+template<typename F> struct foo<F> { };	// { dg-error "redeclared as" }

	Jakub

Re: [C++ PATCH] Avoid ice-on-invalid in xref_basetypes (PR c++/34089, take 2)

[Mark Mitchell]

Jakub Jelinek wrote:

> 2007-11-20  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR c++/34089
> 	* parser.c (cp_parser_class_head): Reject function template ids.
> 
> 	* g++.dg/template/crash74.C: New test.

Yes, that looks good.

Thanks,

-- 
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713