[Bug c/16330] New: Unions handled incorrectly with -fshort-enums

[Bug c/16330] New: Unions handled incorrectly with -fshort-enums

[miko at stacken dot kth dot se]

The following code correctly prints 1 when compiled without -fshort-enums, but 
incorrectly prints a 0 when compiled with -fshort-enums. I assume this is an
endian problem as it works correctly on a i686-pc-linux-gnu target.


#include <stdio.h>

enum E { a, b };

union Y
{
  int x;
  enum E   y;
  void* z;
};

int main()
{
  union Y un;
  un.x=1;
  printf("%d\n", un.y);
}

-- 
           Summary: Unions handled incorrectly with -fshort-enums
           Product: gcc
           Version: 3.3.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: miko at stacken dot kth dot se
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: sparc-sun-solaris
  GCC host triplet: sparc-sun-solaris
GCC target triplet: sparc-sun-solaris


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16330

[Bug c/16330] Unions handled incorrectly with -fshort-enums

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-07-02 16:50 -------
Not a bug, the code (with -fshort-enum) is equivent to:
#include <stdio.h>

enum E { a, b };

union Y
{
  int x;
  short   y;
  void* z;
};

int main()
{
  union Y un;
  un.x=1;
  printf("%d\n", un.y);
}

Again this is not a bug.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16330

[Bug c/16330] Unions handled incorrectly with -fshort-enums

[miko at stacken dot kth dot se]

------- Additional Comments From miko at stacken dot kth dot se  2004-07-02 17:02 -------
You're going to have to explain this one to me.

1) It works as expected on other platforms
2) If I use char or short, I get 0 in both cases. At least this is consistent,
even though it's surely wrong? But by using an enum, which under -fshort-enums
should be resolved to a char, not a short (single byte is the smallest that can
contain all of the values, and is how large gcc allocates in a struct) it
produces two different values???

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16330

[Bug c/16330] Unions handled incorrectly with -fshort-enums

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-07-02 17:18 -------
Basically it is because of endian differences, read a good book about endian to
get the full explaination.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16330

[Bug c/16330] Unions handled incorrectly with -fshort-enums

[miko at stacken dot kth dot se]

------- Additional Comments From miko at stacken dot kth dot se  2004-07-02 17:37 -------
I understand endian differences. For some reason I assumed the compiler would
take care of that for me, by storing the enum in [offset+3] rather than in
[offset+0]. One of the downsides to porting I suppose...

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16330