[Bug c++/61987] New: Name Resolution within a Template

[Bug c++/61987] New: Name Resolution within a Template

[gaoyuanming at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

            Bug ID: 61987
           Summary: Name Resolution within a Template
           Product: gcc
           Version: 4.8.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gaoyuanming at hotmail dot com

Created attachment 33222
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33222&action=edit
Assembly code

I compiled the piece of code:

#include <stdlib.h>

int foo(double d) { return static_cast<int>(d); }

template <typename T>
class Test
{
public:
        Test(int id, T const & value) : m_id(id), m_value(value) {}

        void change()
        {
                m_value = foo(m_id);
        }

        void change(int seed)
        {
                m_id = foo(m_value) * seed;
        }

        T value() const { return m_value; }
private:
        int m_id;
        T   m_value;
};

int foo(int n) { return n; }

int main()
{
    Test<int> t(10, 20);
    t.change();
    t.change(2);

    Test<unsigned long> tul(10, 30);
    tul.change();
    tul.change(3);

    return 0;
}

and found that the function void Test<int>::change(int) calls int foo(double)
to implement it. It is a bug.

[Bug c++/61987] Name Resolution within a Template

[gaoyuanming at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

--- Comment #1 from Yuanming Gao <gaoyuanming at hotmail dot com> ---
Created attachment 33223
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33223&action=edit
source code

[Bug c++/61987] Name Resolution within a Template

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I don't think this is a bug as foundmental types does not have an associated
namespace associated with it. So the overload set is only what is declared
before the template.

[Bug c++/61987] Name Resolution within a Template

[gaoyuanming at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

--- Comment #3 from Yuanming Gao <gaoyuanming at hotmail dot com> ---
Please read the book: <<Inside the Cpp Object Mode>>. The code comes from <<7.1
Templates>>. The author thought the function int foo(int) must be selected.


Best regards,
Yuanming Gao

> From: gcc-bugzilla@gcc.gnu.org
> To: gaoyuanming@hotmail.com
> Subject: [Bug c++/61987] Name Resolution within a Template
> Date: Fri, 1 Aug 2014 09:46:21 +0000
> 
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987
> 
> --- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
> I don't think this is a bug as foundmental types does not have an associated
> namespace associated with it. So the overload set is only what is declared
> before the template.
> 
> -- 
> You are receiving this mail because:
> You reported the bug.

[Bug c++/61987] Name Resolution within a Template

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Yuanming Gao from comment #3)
> Please read the book: <<Inside the Cpp Object Mode>>. The code comes from
> <<7.1 Templates>>. The author thought the function int foo(int) must be
> selected.

Well I disagree with that book based on what the c++ standard says about
foundmental types and how argument dependent lookup works.


>  
>  
> Best regards,
> Yuanming Gao
>  
> > From: gcc-bugzilla@gcc.gnu.org
> > To: gaoyuanming@hotmail.com
> > Subject: [Bug c++/61987] Name Resolution within a Template
> > Date: Fri, 1 Aug 2014 09:46:21 +0000
> > 
> > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987
> > 
> > --- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
> > I don't think this is a bug as foundmental types does not have an associated
> > namespace associated with it. So the overload set is only what is declared
> > before the template.
> > 
> > -- 
> > You are receiving this mail because:
> > You reported the bug.

[Bug c++/61987] Name Resolution within a Template

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
G++ and clang++ call f(double) four times, because that is the only function
visible in the template definition. foo(int) is only visible at the point of
instantiation, so could only be found by ADL, but int has no associated
namespaces.

EDG does call foo(int) twice, and fails to compile because the call with
unsigned long argument could call foo(double) or foo(int) and so is ambiguous.

So either the code never calls foo(int) or the code doesn't compile!

[Bug c++/61987] Name Resolution within a Template

[gaoyuanming at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

--- Comment #6 from Yuanming Gao <gaoyuanming at hotmail dot com> ---
Visual C++ can select the  int foo(int) correctly. I don't know whether it is
an implementation issue, or it is by design.

Best regards,
Yuanming

> From: gcc-bugzilla@gcc.gnu.org
> To: gaoyuanming@hotmail.com
> Subject: [Bug c++/61987] Name Resolution within a Template
> Date: Fri, 1 Aug 2014 10:06:00 +0000
> 
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987
> 
> --- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
> (In reply to Yuanming Gao from comment #3)
> > Please read the book: <<Inside the Cpp Object Mode>>. The code comes from
> > <<7.1 Templates>>. The author thought the function int foo(int) must be
> > selected.
> 
> Well I disagree with that book based on what the c++ standard says about
> foundmental types and how argument dependent lookup works.
> 
> 
> >  
> >  
> > Best regards,
> > Yuanming Gao
> >  
> > > From: gcc-bugzilla@gcc.gnu.org
> > > To: gaoyuanming@hotmail.com
> > > Subject: [Bug c++/61987] Name Resolution within a Template
> > > Date: Fri, 1 Aug 2014 09:46:21 +0000
> > > 
> > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987
> > > 
> > > --- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
> > > I don't think this is a bug as foundmental types does not have an associated
> > > namespace associated with it. So the overload set is only what is declared
> > > before the template.
> > > 
> > > -- 
> > > You are receiving this mail because:
> > > You reported the bug.
> 
> -- 
> You are receiving this mail because:
> You reported the bug.

[Bug c++/61987] Name Resolution within a Template

[gaoyuanming at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

--- Comment #7 from Yuanming Gao <gaoyuanming at hotmail dot com> ---
Please read the book: <<Inside the Cpp Object Mode>>. The code comes from <<7.1
Templates>>. The author thought the function int foo(int) must be selected.

Visual C++ can select the int foo(int) correctly. I don't know whether it is an
implementation issue, or it is by design.

Best regards,
Yuanming

> From: gcc-bugzilla@gcc.gnu.org
> To: gaoyuanming@hotmail.com
> Subject: [Bug c++/61987] Name Resolution within a Template
> Date: Fri, 1 Aug 2014 10:28:36 +0000
> 
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987
> 
> --- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
> G++ and clang++ call f(double) four times, because that is the only function
> visible in the template definition. foo(int) is only visible at the point of
> instantiation, so could only be found by ADL, but int has no associated
> namespaces.
> 
> EDG does call foo(int) twice, and fails to compile because the call with
> unsigned long argument could call foo(double) or foo(int) and so is ambiguous.
> 
> So either the code never calls foo(int) or the code doesn't compile!
> 
> -- 
> You are receiving this mail because:
> You reported the bug.

[Bug c++/61987] Name Resolution within a Template

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

--- Comment #8 from Jonathan Wakely <redi at gcc dot gnu.org> ---
VC++ and the book are both wrong.

[Bug c++/61987] Name Resolution within a Template

[daniel.kruegler at googlemail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |daniel.kruegler@googlemail.
                   |                            |com

--- Comment #9 from Daniel Krügler <daniel.kruegler at googlemail dot com> ---
(In reply to Yuanming Gao from comment #7)
> Please read the book: <<Inside the Cpp Object Mode>>. The code comes from
> <<7.1 Templates>>. The author thought the function int foo(int) must be
> selected.
> 
> Visual C++ can select the int foo(int) correctly. I don't know whether it is
> an implementation issue, or it is by design.

"Can select" is irrelevant in the context of correctness. As others have
already said: The standard requires to consider *only* the overload
foo(double), because for fundamental types the associated set of namespaces is
empty (see [basic.lookup.argdep] p2 b1). In this case, argument-depending
lookup doesn't happen, and only the overloads that can be found lexically
before the call location as part of the (reduced) unqualified lookup are
considered.
>From gcc-bugs-return-457584-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Fri Aug 01 18:45:31 2014
Return-Path: <gcc-bugs-return-457584-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 6912 invoked by alias); 1 Aug 2014 18:45:28 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 6830 invoked by uid 48); 1 Aug 2014 18:45:22 -0000
From: "jason at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/53492] [4.8 Regression] ICE in retrieve_specialization, at cp/pt.c:985
Date: Fri, 01 Aug 2014 18:45:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.6.1
X-Bugzilla-Keywords: accepts-invalid, ice-on-invalid-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: jason at gcc dot gnu.org
X-Bugzilla-Status: RESOLVED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: jason at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 4.8.4
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: bug_status resolution
Message-ID: <bug-53492-4-PlzWU3Z3Xe@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-53492-4@http.gcc.gnu.org/bugzilla/>
References: <bug-53492-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-08/txt/msg00081.txt.bz2
Content-length: 460

https://gcc.gnu.org/bugzilla/show_bug.cgi?idS492

Jason Merrill <jason at gcc dot gnu.org> changed:

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

--- Comment #7 from Jason Merrill <jason at gcc dot gnu.org> ---
Not backporting fixes for invalid code bugs.

[Bug c++/61987] Name Resolution within a Template

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61987

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |INVALID

--- Comment #10 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Closing - GCC does the right thing.