Why does Borland compile with multiple definitions of same object in different C files while GCC does not?

Question

I am studying the behavior of global variables.

So far , I thought the multiple definition of global variables is an illegal way , and must get an error. But I got an unexpected result from Borland C/C++ compiler , while GCC gave me the expected result.

Code:

test1.c:

#include<stdio.h>

void func(void);

int num=1;

void main(){
    func();
    return;
}

test2.c:

#include<stdio.h>

int num=2;

void func(){    
    printf("%d",num);
    return;
}

On MS-DOS prompt

• Borland C/C++ :

  c:\test>bcc32 test1.c test2.c
  

• GCC :

  c:\test>gcc test1.c test2.c
  

Results

• Borland C/C++ :

There's no error and compile&link successfully(This is unexpected for me).After executing test1.exe , 2 was printed on the console. This is num's value defined in test2.c.

• GCC :

GCC gave me an error of multiple definition of num. Of course , a.exe was not made.(This is what I was expecting)

Why does that happen? Please let me know. Thank you!

Answer 1

Multiple external definitions of an object is undefined behavior in C. A common extension is to accept multiple definitions if they don't disagree (usually with same type and no initialization value).

C99 6.9p5 says:

If an identifier declared with external linkage is used in an expression (other than as part of the operand of a sizeof operator whose result is an integer constant), somewhere in the entire program there shall be exactly one external definition for the identifier; otherwise, there shall be no more than one"

and C99, 4.p2:

violation of a "shall" outside of a constraint implies UB