↓Skip to main content
  1. System Level/

Not My Type

·2 mins
C Types Variables Undefined Behavior Pointer Math

In a previous post, I walked through a few basic examples of type systems in managed languages.

The below is a quick type example from the C programming language.

include <stdio.h>

int main(void)
{ 
    int y = 5 + "2"; /* Line 5 */
    printf("%d\n", y); /* Line 6 */
    char *x = 3 + "2"; /* Line 7 */
    printf("%d\n", x); /* Line 8 */
}

/*  Compiler warnings
    The compiler issues the below warnings (not errors)
    but it compiles the program and it can be executed 

    "warning: adding 'int' to a string does 
    not append to the string [-Wstring-plus-int]" Line 5 

    "use array indexing to silence this warning" Line 5

    "warning: incompatible pointer to integer conversion initializing 
    'int' with an expression of type 'char *' [-Wint-conversion]" Line 5 

    "warning: adding 'int' to a string does not append 
    to the string [-Wstring-plus-int] Line 7

    "use array indexing to silence this warning" Line 7 

    "warning: format specifies type 'int' but the argument 
    has type 'char *' [-Wformat]" Line 8  */


/*  Output: (Undefined Behavior) 
    170446759
    170446757
*/

“2” is a string literal. In C, a string literal is stored in memory (allocated by the compiler i.e. “statically allocated”) as an array of characters. Each character in the array has a size of one byte.

C interprets line 5 and line 7 as an attempt to create a pointer (y and x) that reference an address in memory that is 5 bytes (line 5) or 3 bytes (line 7) beyond the end of the string “2”.

C interprets the “+” as pointer math involving the number operand on the left and the size (a character is one byte), and starting address operand on the right (“2”). So 5 times one byte added on to the end of “2”’s location, or 3 times one byte added on to the end of “2”’s location.

In summary:

  • C is statically and weakly typed.
    • The compiler provides type warnings but not type errors.
  • When “integer” + “string”
    • C does not attempt to implicitly convert the integer to a string or the string to an integer
    • Instead C views the statement as an attempt to create a new pointer to a storage location with a size and address that is a function of the integer value and the string’s size and location.
    • This results in undefined behavior
↑