[Bug libstdc++/54249] New: [C++11] No ::nullptr_t in header <stddef.h>

[Bug libstdc++/54249] New: [C++11] No ::nullptr_t in header <stddef.h>

[daniel.kruegler at googlemail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

             Bug #: 54249
           Summary: [C++11] No ::nullptr_t in header <stddef.h>
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: daniel.kruegler@googlemail.com


gcc 4.8.0 20120729 (experimental) rejects the following code:

#include <stddef.h>

::nullptr_t n;

"3|error: 'nullptr_t' in namespace '::' does not name a type"

This code is supposed to be accepted, because [depr.c.headers] p2 says:

"Every C header, each of which has a name of the form name.h, behaves as if
each name placed in the standard library namespace by the corresponding cname
header is placed within the global namespace scope."

Paolo Carlini has suggested to perform this change near to existing C++11-aware
code in stddef.h:

#if (defined (__STDC_VERSION__) && __STDC_VERSION__ >= 201112L) \
  || (defined(__cplusplus) && __cplusplus >= 201103L)
#ifndef _GCC_MAX_ALIGN_T
#define _GCC_MAX_ALIGN_T
/* Type whose alignment is supported in every context and is at least
   as great as that of any standard type not using alignment
   specifiers.  */
typedef struct {
  long long __max_align_ll __attribute__((__aligned__(__alignof__(long
long))));
  long double __max_align_ld __attribute__((__aligned__(__alignof__(long
double))));
} max_align_t;
#endif
#endif /* C11 or C++11.  */

[I guess I have selected the wrong component. Please correct, if necessary]

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[daniel.kruegler at googlemail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

Daniel Krügler <daniel.kruegler at googlemail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |rejects-valid

--- Comment #1 from Daniel Krügler <daniel.kruegler at googlemail dot com> 2012-08-14 08:05:05 UTC ---
I should add that the code example was compiled in C++11 mode

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[glisse at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

Marc Glisse <glisse at gcc dot gnu.org> changed:

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

--- Comment #2 from Marc Glisse <glisse at gcc dot gnu.org> 2012-08-14 09:36:15 UTC ---
(In reply to comment #0)
> "Every C header, each of which has a name of the form name.h, behaves as if
> each name placed in the standard library namespace by the corresponding cname
> header is placed within the global namespace scope."

Note that libstdc++ has this wrong for almost every header where the C++
standard deviates from the C standard (there is already a PR or two about that
somewhere). stddef.h is an exception, since it is provided by gcc, it can
indeed be fixed quite easily :-)

I believe someone (Howard maybe?) wanted to file a DR asking to remove this
paragraph.

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[daniel.kruegler at googlemail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #3 from Daniel Krügler <daniel.kruegler at googlemail dot com> 2012-08-14 09:52:30 UTC ---
(In reply to comment #2)
> (In reply to comment #0)
> > "Every C header, each of which has a name of the form name.h, behaves as if
> > each name placed in the standard library namespace by the corresponding cname
> > header is placed within the global namespace scope."

[..] 

> I believe someone (Howard maybe?) wanted to file a DR asking to remove this
> paragraph.

I don't want to start a lengthy discussion here about the C++ Standard Library
specification, but it must be clear that removing above paragraph would have
the effect of making the effects of including any <*.h> header in a C++ program
undefined, because [depr.c.headers] is the only place in the Library
specification that defines the semantics.

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[glisse at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #4 from Marc Glisse <glisse at gcc dot gnu.org> 2012-08-14 10:35:56 UTC ---
(In reply to comment #3)
> I don't want to start a lengthy discussion here about the C++ Standard Library
> specification, but it must be clear that removing above paragraph would have
> the effect of making the effects of including any <*.h> header in a C++ program
> undefined, because [depr.c.headers] is the only place in the Library
> specification that defines the semantics.

Oups, "amend" instead of "remove". But you are right, this is not the place, I
just wanted to mention the existence of a debate about the .h headers.

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #5 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-08-14 10:38:45 UTC ---
It's true however, that, as I mentioned already somewhere, in general our
implementation doesn't have control over the underlying *.h headers. Thus, it
seems mildly inconsistent to add things only to the *few* *.h headers over
which we do have control. And, indeed, apparently the Standard is in some sort
of half way state, because LWG already morally acknowledged that libstdc++-v3
is not special about this with the resolution of LWG 456, which allows the c*
headers to define entities both in the global namespace and in std:: exactly
because (right? Why otherwise?) it's now allowed for a c* to just include
whatever *.h it finds on the system.

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[glisse at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #6 from Marc Glisse <glisse at gcc dot gnu.org> 2012-08-14 10:48:48 UTC ---
(In reply to comment #5)
> It's true however, that, as I mentioned already somewhere, in general our
> implementation doesn't have control over the underlying *.h headers.

Although it does have control over potential *.h headers it ships and puts
first in the include path (as it does for complex.h and fenv.h for instance).

> Thus, it seems mildly inconsistent to add things only to the *few* *.h
> headers over which we do have control.

Grab the low-hanging fruits?

(sorry Daniel, conversations seems to have a way to start when we don't want
them to...)

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-08-14 10:50:40 UTC ---
(In reply to comment #5)
> seems mildly inconsistent to add things only to the *few* *.h headers over
> which we do have control.

Better to be mildly inconsistent than ship a knowingly non-conformant stddef.h

If we controlled all the headers we'd fix them, so we should fix the ones we do
control (and hope that the native libc is or will eventually be C++ friendly)

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #8 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-08-14 10:52:38 UTC ---
Sure, grab the low-hanging fruits, like this one ;) But while we do that we
should also do an audit and establish what is missing from headers which we do
not control and at least talk to the glibc people about those. Assuming the
Standard doesn't change of course ;)

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #9 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-08-14 10:56:02 UTC ---
I must also add that frankly personally I don't consider these issues so
serious: IMHO a new C++ project should immediately use the c* headers. I see
these issues mostly about legacy code.

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #10 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-08-14 10:59:41 UTC ---
N.B. see also http://gcc.gnu.org/ml/gcc-patches/2012-01/msg00375.html and
http://gcc.gnu.org/ml/gcc-patches/2012-02/msg00192.html and the stdalign.h part
of http://gcc.gnu.org/ml/gcc-patches/2011-11/msg00893.html which currently
cause non-conformance even when including <cstdbool> or <cstdalign>

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[daniel.kruegler at googlemail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #11 from Daniel Krügler <daniel.kruegler at googlemail dot com> 2012-08-14 11:01:47 UTC ---
(In reply to comment #9)
> I must also add that frankly personally I don't consider these issues so
> serious: IMHO a new C++ project should immediately use the c* headers. I see
> these issues mostly about legacy code.

Completely agreed and therefore reducing the priority. The main reason for this
bugzilla entry was not, because I need above code to be well-formed, but
because there often exists uncertainty in the wild about the interpretation of
the semantics of including the <*.h> headers. The issue came out of a lengthy
discussion that I had about this semantics on the std-discussion group.

[Bug libstdc++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[daniel.kruegler at googlemail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

Daniel Krügler <daniel.kruegler at googlemail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |minor

[Bug c++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2012-09-28
          Component|libstdc++                   |c++
         AssignedTo|unassigned at gcc dot       |paolo.carlini at oracle dot
                   |gnu.org                     |com
   Target Milestone|---                         |4.8.0
     Ever Confirmed|0                           |1

--- Comment #12 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-09-28 10:00:36 UTC ---
On it.

[Bug c++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

[Bug c++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[paolo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

--- Comment #13 from paolo at gcc dot gnu.org <paolo at gcc dot gnu.org> 2012-10-06 22:44:23 UTC ---
Author: paolo
Date: Sat Oct  6 22:44:12 2012
New Revision: 192173

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=192173
Log:
2012-10-06  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/54249
    * ginclude/stddef.h: In C++11 mode declare nullptr_t in the global
    namespace.

/testsuite
2012-10-06  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/54249
    * g++.dg/cpp0x/stddef.C: New.

Added:
    trunk/gcc/testsuite/g++.dg/cpp0x/stddef.C
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ginclude/stddef.h
    trunk/gcc/testsuite/ChangeLog

[Bug c++/54249] [C++11] No ::nullptr_t in header <stddef.h>

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54249

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #14 from Paolo Carlini <paolo.carlini at oracle dot com> 2012-10-06 22:46:39 UTC ---
Fixed.