🚀 go-pugleaf

tanmoy@qcd.lanl.gov (Tanmoy Bhattacharya) writes:

>In article <4elq8a$j0b@taurus.fccc.edu> slifker@castor.fccc.edu
>(Michael J. Slifker) writes: 
>
>   I've compiled the following program with gcc:
>
>   -----------
>
>   #include <stdio.h>
>
>   char *str1 = "Hello, World\n";
>   const char *str2 = "Hello, World\n";
>
>   int main(void)
>   {
>     str1[0] = 'J';
>     printf("%s", str2);
>     return 0;
>   }
>
>   -----------
>
>   ...and it prints out "Jello, World". This seems odd, at least to
>   me. Is this a bug, or is the behavior above officially undefined?
>
>Attempt to modify a string literal results in undefined behaviout. Any
>behaviour by the implementation is then valid.

The full answer is in ANSI classic 3.1.4:

    Identical string literals of either form need not be distinct.  If
    the program attempts to modify a string literal of either form, the
    behavior is undefined.

So, if we modify main() in the above example to look like this:

    int main(void)
    {
       printf("%d\n", str1 == str2);
       return 0;
    }

Both "0" and "1" are correct outputs.  Different compilers on the same
platform may produce diferent results:

    ues5:~/tmp 88> gcc test.c
    ues5:~/tmp 89> ./a.out 
    1
    ues5:~/tmp 90> cc test.c
    ues5:~/tmp 91> ./a.out
    0

This actually explains the (undefined) behaviour observed by the original
poster.  There are systems where the attempt to modify a string literal
results in a segmentation fault, because string literals are put in a 
read-only memory segment.

Dan
-- 
Dan Pop
CERN, CN Division
Email: danpop@mail.cern.ch 
Mail:  CERN - PPE, Bat. 31 R-004, CH-1211 Geneve 23, Switzerland