[Bug c++/51424] New: [C++11] G++ should diagnose self-delegating constructors

[Bug c++/51424] New: [C++11] G++ should diagnose self-delegating constructors

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51424

             Bug #: 51424
           Summary: [C++11] G++ should diagnose self-delegating
                    constructors
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Keywords: diagnostic
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: jason@gcc.gnu.org
                CC: pedro.lamarao@gmail.com, ville.voutilainen@gmail.com


This modification of g++.dg/template/meminit1.C should get a diagnostic in
C++11 mode as well:

template <class T >
struct S
{
  S() : S() {} // { dg-message "delegating constructors" "" { target c++98 } }
  // { dg-error "delegates to itself" "" { target c++11 } }
};

S<int> s;

because 12.6.2 says,

"If a constructor delegates to itself directly or indirectly, the program is
ill-formed; no diagnostic is required."

The diagnostic is not required, but it is easy to give, so we should do so.

[Bug c++/51424] [C++11] G++ should diagnose self-delegating constructors

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51424

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2013-08-29
           Assignee|unassigned at gcc dot gnu.org      |paolo.carlini at oracle dot com
   Target Milestone|---                         |4.9.0
     Ever confirmed|0                           |1

--- Comment #2 from Paolo Carlini <paolo.carlini at oracle dot com> ---
I have a very simple patch which appears to work fine (for the very simple
cases ;) but constexpr constructors, virtual inheritance, etc are included)

Index: init.c
===================================================================
--- init.c    (revision 202070)
+++ init.c    (working copy)
@@ -500,8 +500,9 @@ perform_target_ctor (tree init)
   tree decl = current_class_ref;
   tree type = current_class_type;

-  finish_expr_stmt (build_aggr_init (decl, init, LOOKUP_NORMAL,
-                                     tf_warning_or_error));
+  finish_expr_stmt (build_aggr_init (decl, init,
+                     LOOKUP_NORMAL|LOOKUP_DELEGATING_CONS,
+                     tf_warning_or_error));
   if (TYPE_HAS_NONTRIVIAL_DESTRUCTOR (type))
     {
       tree expr = build_delete (type, decl, sfk_complete_destructor,
@@ -1667,6 +1668,13 @@ expand_default_init (tree binfo, tree true_exp, tr
   if (parms != NULL)
     release_tree_vector (parms);

+  if ((complain & tf_error)
+      && (flags & LOOKUP_DELEGATING_CONS)
+      && TREE_CODE (rval) == CALL_EXPR
+      && (DECL_ABSTRACT_ORIGIN (TREE_OPERAND (CALL_EXPR_FN (rval), 0))
+      == current_function_decl))
+    error ("constructor delegates to itself");
+
   if (exp == true_exp && TREE_CODE (rval) == CALL_EXPR)
     {
       tree fn = get_callee_fndecl (rval);
Index: cp-tree.h
===================================================================
--- cp-tree.h    (revision 202070)
+++ cp-tree.h    (working copy)
@@ -4509,6 +4509,8 @@ enum overload_flags { NO_SPECIAL = 0, DTOR_FLAG, T
 #define LOOKUP_NO_RVAL_BIND (LOOKUP_EXPLICIT_TMPL_ARGS << 1)
 /* Used by case_conversion to disregard non-integral conversions.  */
 #define LOOKUP_NO_NON_INTEGRAL (LOOKUP_NO_RVAL_BIND << 1)
+/* Used for delegating constructors in order to diagnose self-delegation.  */
+#define LOOKUP_DELEGATING_CONS (LOOKUP_NO_NON_INTEGRAL << 1)

 #define LOOKUP_NAMESPACES_ONLY(F)  \
   (((F) & LOOKUP_PREFER_NAMESPACES) && !((F) & LOOKUP_PREFER_TYPES))

[Bug c++/51424] [C++11] G++ should diagnose self-delegating constructors

[paolo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51424

--- Comment #3 from paolo at gcc dot gnu.org <paolo at gcc dot gnu.org> ---
Author: paolo
Date: Fri Aug 30 15:39:01 2013
New Revision: 202110

URL: http://gcc.gnu.org/viewcvs?rev=202110&root=gcc&view=rev
Log:
/cp
2013-08-30  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/51424
    * cp-tree.h (LOOKUP_DELEGATING_CONS): Add.
    * init.c (perform_target_ctor): Use it.
    * call.c (build_special_member_call): Diagnose self-delegating
    constructors.

/testsuite
2013-08-30  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/51424
    * g++.dg/cpp0x/dc8.C: New.
    * g++.dg/template/meminit1.C: Adjust.

Added:
    trunk/gcc/testsuite/g++.dg/cpp0x/dc8.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/call.c
    trunk/gcc/cp/cp-tree.h
    trunk/gcc/cp/init.c
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/g++.dg/template/meminit1.C

[Bug c++/51424] [C++11] G++ should diagnose self-delegating constructors

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51424

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.9.0                       |4.9.1

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 4.9.0 has been released

[Bug c++/51424] [C++11] G++ should diagnose self-delegating constructors

[ville.voutilainen at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51424

--- Comment #5 from Ville Voutilainen <ville.voutilainen at gmail dot com> ---
(In reply to Jakub Jelinek from comment #4)
> GCC 4.9.0 has been released

Well, shouldn't this bug be closed, then? The patch Paolo wrote was applied and
shipped in 4.9.0, so this is already in, no need to change the milestone to
4.9.1.

[Bug c++/51424] [C++11] G++ should diagnose self-delegating constructors

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51424

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
                 CC|                            |jakub at gcc dot gnu.org
         Resolution|---                         |FIXED
   Target Milestone|4.9.1                       |4.9.0

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Then it should have been closed months ago.

[Bug c++/51424] [C++11] G++ should diagnose self-delegating constructors

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51424

--- Comment #7 from Paolo Carlini <paolo.carlini at oracle dot com> ---
I seem to remember that the reason why I didn't close it completely at the time
is that we may want to extend the diagnostic to cover non-trivial cycles too...
Opinions? Could also be a separate bug, that for sure.

[Bug c++/51424] [C++11] G++ should diagnose self-delegating constructors

[ville.voutilainen at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51424

--- Comment #8 from Ville Voutilainen <ville.voutilainen at gmail dot com> ---
My 2 cents is to open a separate bug if we want diagnostics for the more
elaborate self-delegating cases.