Incompatible binutils and gcc

[Sergey E. Gorelkin]

Hi All,

I am install two Linux Redhat 6.2 and 7.1 on i386 machine (host == target).
At RedHat 6.2:
binutils-2.9.5.0.22
gcc-2.95.3

At RedHat 7.1:
binutils-2.10.91.0.2
gcc-2.96-81

I wrote some test.cpp file for example:
class A {
public:
  int x;
  double y;
};

main() {
  A a;
  a.x = 1;
  a.y = 1.5;
}

I am compile with gcc version of 2.95 and run
g++ -g test.cpp
objdump -g a.out
This all right.

But if i try use gcc version 2.96 or hider
g++ -g test.cpp
objdump -g a.out
Failed with :
a.out:     file format elf32-i386
bad mangled name `1A__as__1ARC1A'
Last stabs entries before error:
n_type n_desc n_value  string
LSYM   0      00000000 short int:t(0,8)address@hidden;r(0,8);-32768;32767;
LSYM   0      00000000 short unsigned int:t(0,9)address@hidden;r(0,9);0;65535;
LSYM   0      00000000 signed char:t(0,10)address@hidden;r(0,10);-128;127;
LSYM   0      00000000 unsigned char:t(0,11)address@hidden;r(0,11);0;255;
LSYM   0      00000000 bool:t(0,12)address@hidden;-16;
LSYM   0      00000000 float:t(0,13)=r(0,1);4;0;
LSYM   0      00000000 double:t(0,14)=r(0,1);8;0;
LSYM   0      00000000 long double:t(0,15)=r(0,1);12;0;
LSYM   0      00000000 complex 
int:t(0,16)=s8real:(0,1),0,32;imag:(0,1),32,32;;
LSYM   0      00000000 complex float:t(0,17)=r(0,17);8;0;
LSYM   0      00000000 complex double:t(0,18)=r(0,18);16;0;
LSYM   0      00000000 complex long double:t(0,19)=r(0,19);24;0;
LSYM   0      00000000 __builtin_va_list:t(0,20)=*(0,21)=(0,21)
LSYM   0      00000000 __wchar_t:t(0,22)=r(0,22);-2147483648;2147483647;
LSYM   0      00000000 __vtbl_ptr_type:t(0,23)=*(0,24)=f(0,1)
LSYM   0      00000000 
A:T(0,25)=s16next:/0(0,26)=*(0,25),0,32;x:(0,1),32,32;y:(0,14),64,64;operator=::(0,27)=##(0,28)=&(0,25);:__as__1ARC1A;2A.;A::(0,29)=##(0,26);:RC1A;2A.(0,30)=##(0,26);:;2A.;; 


Best Regards,
Sergey.