[Bug c++/11374] New: Call by reference using enum casted to int passes in wrong address

[mike at mikesclutter dot com]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11374

           Summary: Call by reference using enum casted to int passes in
                    wrong address
           Product: gcc
           Version: 3.3
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: mike at mikesclutter dot com
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: ?
  GCC host triplet: ?
GCC target triplet: ?

Platform: Linux 2.4.20 #3 Fri May 30 08:25:22 CDT 2003 i686 Intel(R)   
          Pentium(R) III Mobile CPU 1133MHz GenuineIntel GNU/Linux

Overview:

This problem did not occur in previous versions of gcc (on linux at least).
Basically, call by reference when using enum arguments (casted as int) does not
provide the correct reference to the called function.

A simple example:

// The enum
enum _t
{
    off = 0,
    on = 100
};

// The function
void fn(int& x)
{
    // The rub: &x != &t in main(), although it should.
    printf("&x in fn(): %x\n", &x);

    x = on;

    printf("value set in fn(): %i\n", x);
}

// The caller
void main()
{
    _t t = off;

    printf("&t in main(): %x\n", &t);

    fn((int)t);

    printf("value received in main(): %i\n", t);
}

This produces the following output:

&t in main(): bffff6d0
&x in fn(): bffff6cc
value set in fn(): 100
value received in main(): 0

Thus, since &t != &x, the variable t is never set to the proper value by fn()
even though it is called by reference.