[Bug c++/48008] New: [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[Bug c++/48008] New: [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[zsojka at seznam dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48008

           Summary: [4.6 Regression]
                    libstdc++-v3/testsuite/20_util/function/43397.cc fails
                    to assemble with -fabi-version=5
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: zsojka@seznam.cz


I am not sure if this is a bug, but as there were commits to the ABI code, and
the release is probably near, I opened this PR without more analysing... I am
sorry if it renders invalid.

This is the compiler output:
$ g++ testcase2.C -std=c++0x -fabi-version=5 -c
/tmp/cclBTCq8.s: Assembler messages:
/tmp/cclBTCq8.s:314: Error: symbol
`_ZSt6mem_fnIFiOiE3FooESt7_Mem_fnIMT0_T_ES6_' is already defined
/tmp/cclBTCq8.s:352: Error: symbol
`_ZSt6mem_fnIFiOiE3FooESt7_Mem_fnIMT0_T_ES6_' is already defined
/tmp/cclBTCq8.s:390: Error: symbol
`_ZSt6mem_fnIFiOiE3FooESt7_Mem_fnIMT0_T_ES6_' is already defined

$ echo _ZSt6mem_fnIFiOiE3FooESt7_Mem_fnIMT0_T_ES6_ | c++filt
std::_Mem_fn<int (Foo::*)(int&&)> std::mem_fn<int ()(int&&), Foo>(int
(Foo::*)(int&&))


With -fabi-version=1 (or gcc 4.5), the following symbols are generated:
$ echo _ZSt6mem_fnIFiOiE3FooESt7_Mem_fnIMT0_T_ES6_ | c++filt
std::_Mem_fn<int (Foo::*)(int&&)> std::mem_fn<int ()(int&&), Foo>(int
(Foo::*)(int&&))
$ echo _ZSt6mem_fnIKFiOiE3FooESt7_Mem_fnIMT0_T_ES7_ | c++filt
std::_Mem_fn<int ( const Foo::*)(int&&)> std::mem_fn<int ( const)(int&&),
Foo>(int ( const Foo::*)(int&&))
$ echo _ZSt6mem_fnIVFiOiE3FooESt7_Mem_fnIMT0_T_ES7_ | c++filt
std::_Mem_fn<int ( volatile Foo::*)(int&&)> std::mem_fn<int ( volatile)(int&&),
Foo>(int ( volatile Foo::*)(int&&))
$ echo _ZSt6mem_fnIVKFiOiE3FooESt7_Mem_fnIMT0_T_ES7_ | c++filt
std::_Mem_fn<int ( const volatile Foo::*)(int&&)> std::mem_fn<int ( const
volatile)(int&&), Foo>(int ( const volatile Foo::*)(int&&))

[Bug c++/48008] [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[zsojka at seznam dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48008

--- Comment #1 from Zdenek Sojka <zsojka at seznam dot cz> 2011-03-06 17:32:33 UTC ---
Created attachment 23562
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23562
reduced testcase

Reduced from libstdc++-v3/testsuite/20_util/function/43397.cc
Also, it compiles with 4.5 (compared to original testcase).

[Bug c++/48008] [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48008

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.0

[Bug c++/48008] [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48008

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011.03.07 10:43:51
     Ever Confirmed|0                           |1

--- Comment #2 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-03-07 10:43:51 UTC ---
Definitely needs investigation before 4.6.0.

Confirmed.

[Bug c++/48008] [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48008

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P1                          |P3
             Status|NEW                         |ASSIGNED
   Target Milestone|4.6.0                       |---

--- Comment #3 from Jason Merrill <jason at gcc dot gnu.org> 2011-03-07 16:58:45 UTC ---
Oh, yuck.  The issue here is that attribute const/noreturn use the same
representation as function cv-quals, so when I fixed the compiler to ignore the
attributes it started ignoring the function cv-quals as well.  Should
definitely fix this for 4.6.

[Bug c++/48008] [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48008

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
   Target Milestone|---                         |4.6.0

[Bug c++/48008] [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48008

--- Comment #4 from Jason Merrill <jason at gcc dot gnu.org> 2011-03-07 20:10:56 UTC ---
Author: jason
Date: Mon Mar  7 20:10:44 2011
New Revision: 170752

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=170752
Log:
    PR c++/48008
    * mangle.c (write_type): Strip cv-quals from FUNCTION_TYPE here.
    (write_CV_qualifiers_for_type): Not here.

Added:
    trunk/gcc/testsuite/g++.dg/abi/mangle46.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/mangle.c
    trunk/gcc/testsuite/ChangeLog

[Bug c++/48008] [4.6 Regression] libstdc++-v3/testsuite/20_util/function/43397.cc fails to assemble with -fabi-version=5

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48008

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

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

--- Comment #5 from Jason Merrill <jason at gcc dot gnu.org> 2011-03-07 20:11:59 UTC ---
Done.