[Bug c++/16726] New: g++ Bug in Linux/PPC with altivec

[Bug c++/16726] New: g++ Bug in Linux/PPC with altivec

[indukuru at us dot ibm dot com]

Test Case which fails :
--------------------------------------------------------------------------------
#include <altivec.h>
class Test
{
        public:
                bool operator==(const Test& in);
                bool operator!=(const Test& in) { return !(*this == in);}
};
---------------------------------------------------------------------------------
Test Case which Passes

class Test
{
        public:
                bool operator==(const Test& in);
                bool operator!=(const Test& in) { return !(*this == in);}
};
------------------------------------------------------------------------------
Output of g++ -v -save-temps 


gpul91:/home/rajeev# gcc-bins-7262004/bin/g++ -v -save-temps -maltivec
-mabi=altivec test1.h
Reading specs from
/home/rajeev/gcc-bins-7262004/lib/gcc/powerpc-unknown-linux-gnu/3.5.0/specs
Configured with: ../gcc/configure --prefix=/home/rajeev/gcc-bins-7262004
--enable-languages=c,c++
Thread model: posix
gcc version 3.5.0 20040726 (experimental)
 /home/rajeev/gcc-bins-7262004/libexec/gcc/powerpc-unknown-linux-gnu/3.5.0/cc1plus
-E -quiet -v -D_GNU_SOURCE -D__unix__ -D_
_gnu_linux__ -D__linux__ -Dunix -D__unix -Dlinux -D__linux -Asystem=linux
-Asystem=unix -Asystem=posix test1.h -maltivec -ma
bi=altivec -fpch-preprocess -o test1.ii
ignoring nonexistent directory
"/home/rajeev/gcc-bins-7262004/lib/gcc/powerpc-unknown-linux-gnu/3.5.0/../../../../powerpc-un
known-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
 /home/rajeev/gcc-bins-7262004/lib/gcc/powerpc-unknown-linux-gnu/3.5.0/../../../../include/c++/3.5.0
 /home/rajeev/gcc-bins-7262004/lib/gcc/powerpc-unknown-linux-gnu/3.5.0/../../../../include/c++/3.5.0/powerpc-unknown-linux-g
nu
 /home/rajeev/gcc-bins-7262004/lib/gcc/powerpc-unknown-linux-gnu/3.5.0/../../../../include/c++/3.5.0/backward
 /usr/local/include
 /home/rajeev/gcc-bins-7262004/include
 /home/rajeev/gcc-bins-7262004/lib/gcc/powerpc-unknown-linux-gnu/3.5.0/include
 /usr/include
End of search list.
 /home/rajeev/gcc-bins-7262004/libexec/gcc/powerpc-unknown-linux-gnu/3.5.0/cc1plus
-fpreprocessed test1.ii -quiet -dumpbase
test1.h -maltivec -mabi=altivec -auxbase test1 -version -o test1.s
--output-pch=test1.h.gch
GNU C++ version 3.5.0 20040726 (experimental) (powerpc-unknown-linux-gnu)
        compiled by GNU C version 3.2.3 20030331 (Debian prerelease).
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
test1.h: In member function `unsigned int __vector__ Test::operator!=(Test*,
const Test&)':
test1.h:9: error: no match for 'operator==' in '*(Test*)this == in'
test1.h:8: note: candidates are: unsigned int __vector__ Test::operator==(Test*,
const Test&)
---------------------------------------------------------------------------------

Target System - Apple G5 running Suse Linux.

-- 
           Summary: g++ Bug in Linux/PPC with altivec
           Product: gcc
           Version: 3.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: indukuru at us dot ibm dot com
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16726

[Bug target/16726] g++ Bug in Linux/PPC with altivec

[bangerth at dealii dot org]

------- Additional Comments From bangerth at dealii dot org  2004-07-26 22:15 -------
Andrew, can you verify this on your PPC? 
 
W. 

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c++                         |target


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16726

[Bug target/16726] g++ Bug in Linux/PPC with altivec

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-07-26 22:18 -------
I was going to confirm it after I build with Dnovillo's patch which fixes my bootstrap problem.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16726

[Bug target/16726] g++ Bug in Linux/PPC with altivec

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-07-26 22:20 -------
But really I think the problem is that altivec.h defines bool as something different than what normal C+
+ expect so it is a bug in the header but really a bug in the PIM rather than anywhere else.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16726

[Bug target/16726] [3.5 Regression] g++ Bug in Linux/PPC with altivec

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-08-15 05:26 -------
Actually this is not a bug, you have to #undef bool after including altivec.h if you want to compile with 
bool as the normal meaning.

This is mainly because we don't have context-sensitive keywords in GCC (except in Apple's GCC).

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
           Keywords|                            |rejects-valid
         Resolution|                            |INVALID
            Summary|g++ Bug in Linux/PPC with   |[3.5 Regression] g++ Bug in
                   |altivec                     |Linux/PPC with altivec


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16726