[Bug c++/65323] New: -Wzero-as-null-pointer-constant stutters

[Bug c++/65323] New: -Wzero-as-null-pointer-constant stutters

[akim.demaille at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

            Bug ID: 65323
           Summary: -Wzero-as-null-pointer-constant stutters
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: akim.demaille at gmail dot com

While tracking a spurious warning about at 0 instead of nullptr, I stumbled on
the following case, where g++ is spitting its warning too many times (4.9 and
5).

struct foo
{
  foo(void* = 0)  {}
  void fun(void* = 0) {}
  void* p = 0;
};

int main()
{
  foo f;
  f.fun();
}


g++-mp-5 -std=c++14 -Wzero-as-null-pointer-constant foo.cc 
foo.cc:3:15: warning: zero as null pointer constant
[-Wzero-as-null-pointer-constant]
   foo(void* = 0)  {}
               ^
foo.cc:3:15: warning: zero as null pointer constant
[-Wzero-as-null-pointer-constant]
foo.cc:4:20: warning: zero as null pointer constant
[-Wzero-as-null-pointer-constant]
   void fun(void* = 0) {}
                    ^
foo.cc:4:20: warning: zero as null pointer constant
[-Wzero-as-null-pointer-constant]
foo.cc:5:13: warning: zero as null pointer constant
[-Wzero-as-null-pointer-constant]
   void* p = 0;
             ^

[Bug c++/65323] -Wzero-as-null-pointer-constant stutters

[paolo.carlini at oracle dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

--- Comment #1 from Paolo Carlini <paolo.carlini at oracle dot com> ---
This one should be easy to fix.

[Bug c++/65323] -Wzero-as-null-pointer-constant stutters

[paolo.carlini at oracle dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |paolo.carlini at oracle dot com

[Bug c++/65323] [4.8/4.9/5 Regression] duplicate -Wzero-as-null-pointer-constant warnings

[paolo.carlini at oracle dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|-Wzero-as-null-pointer-cons |[4.8/4.9/5 Regression]
                   |tant stutters               |duplicate
                   |                            |-Wzero-as-null-pointer-cons
                   |                            |tant warnings

--- Comment #2 from Paolo Carlini <paolo.carlini at oracle dot com> ---
This is in fact a regression which started when check_default_argument got a
perform_implicit_conversion_flags call near the beginning.

[Bug c++/65323] [4.8/4.9/5 Regression] duplicate -Wzero-as-null-pointer-constant warnings

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |diagnostic
           Priority|P3                          |P2
   Target Milestone|---                         |4.8.5

[Bug c++/65323] [4.8/4.9/5 Regression] duplicate -Wzero-as-null-pointer-constant warnings

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

Kai Tietz <ktietz at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ktietz at gcc dot gnu.org

--- Comment #3 from Kai Tietz <ktietz at gcc dot gnu.org> ---
Hmm, the issue seems to be that we emit within check_default_argument () two
times the call to maybe_warn_zero_as_null_pointer_constant ().  Once indirectly
via perform_implicit_conversion_flags () call, and then later on directly
within if-condition for returning nullptr_node.
It seems that second call is superflous in terms of warning.  So I suggest the
following patch:

Index: decl.c
===================================================================
--- decl.c      (Revision 221277)
+++ decl.c      (Arbeitskopie)
@@ -11231,7 +11233,8 @@ check_default_argument (tree decl, tree arg, tsub
       && TYPE_PTR_OR_PTRMEM_P (decl_type)
       && null_ptr_cst_p (arg)
       && (complain & tf_warning)
-      && maybe_warn_zero_as_null_pointer_constant (arg, input_location))
+      && c_inhibit_evaluation_warnings == 0
+      && !NULLPTR_TYPE_P(TREE_TYPE (arg)))
     return nullptr_node;

   /* [dcl.fct.default]

[Bug c++/65323] [4.8/4.9/5 Regression] duplicate -Wzero-as-null-pointer-constant warnings

[paolo.carlini at oracle dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

--- Comment #4 from Paolo Carlini <paolo.carlini at oracle dot com> ---
I sent a patch to the mailing list about this. If we don't want to apply it and
we want to be super-conservative, we can indeed simply do this, with a comment,
in my opinion:

@@ -11227,11 +11243,14 @@
                                     LOOKUP_IMPLICIT);
   --cp_unevaluated_operand;

+  /* FIXME: should be OK to just check
+     TYPE_PTR_OR_PTRMEM_P (decl_type) && null_ptr_cst_p (arg).  */
   if (warn_zero_as_null_pointer_constant
       && TYPE_PTR_OR_PTRMEM_P (decl_type)
       && null_ptr_cst_p (arg)
       && (complain & tf_warning)
-      && maybe_warn_zero_as_null_pointer_constant (arg, input_location))
+      && c_inhibit_evaluation_warnings == 0
+      && !NULLPTR_TYPE_P (TREE_TYPE (arg)))
     return nullptr_node;

   /* [dcl.fct.default]

[Bug c++/65323] [4.8/4.9/5 Regression] duplicate -Wzero-as-null-pointer-constant warnings

[paolo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

--- Comment #5 from paolo at gcc dot gnu.org <paolo at gcc dot gnu.org> ---
Author: paolo
Date: Thu Mar 12 23:55:49 2015
New Revision: 221402

URL: https://gcc.gnu.org/viewcvs?rev=221402&root=gcc&view=rev
Log:
2015-03-12  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/65323
    * decl.c (check_default_argument): Don't call
    maybe_warn_zero_as_null_pointer_constant.

Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/decl.c

[Bug c++/65323] [4.8/4.9 Regression] duplicate -Wzero-as-null-pointer-constant warnings

[paolo.carlini at oracle dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65323

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED
   Target Milestone|4.8.5                       |5.0
            Summary|[4.8/4.9/5 Regression]      |[4.8/4.9 Regression]
                   |duplicate                   |duplicate
                   |-Wzero-as-null-pointer-cons |-Wzero-as-null-pointer-cons
                   |tant warnings               |tant warnings

--- Comment #6 from Paolo Carlini <paolo.carlini at oracle dot com> ---
Fixed for 5.0, not worth fixing in 4.8/4.9.