Mastering Algorithms with C

Function Pointers

Function pointers are pointers that, instead of pointing to data, point to executable code or to blocks of information needed to invoke executable code. They are used to store and manage functions as if they were pieces of data. Function pointers have a type that is described in terms of a return value and parameters that the function accepts. Declarations for function pointers look much like declarations for functions, except that an asterisk ( * ) appears before the function name, and the asterisk and name are surrounded by parentheses for reasons of associativity. For example, in the following code, match is declared as a pointer to a function that accepts two void pointers and returns an integer:

int (*match)(void *key1, void *key2);

This declaration means that we can set match to point to any function that accepts two void pointers and returns an integer. For example, suppose match_int is a function that accepts two void pointers to integers and returns 1 if the integers match, or otherwise. Assuming the previous declaration, we could set match to point to this function by executing the following statement:

match = match_int;

To execute a function referenced by a function pointer, we simply use the function pointer wherever we would normally use the function itself. For example, to invoke the function referenced by match earlier, we execute the following statement, assuming x, y, and retval have been declared as integers:

retval = match(&x, &y);

One important use ...