Thread: defining a function pointer which recieves a function pointer

Just use a couple of typedef's, it makes life so much easier

typedef int (*fooFn)( int );
typedef int (*barFn)( fooFn );

Then
barFn init = 0;
init = GetProcAddress(hInstLibrary, "initialize");
init( aFooFn );

The & is wrong imo.... You're dealing with a pointer, which holds an address. There's no reason to try to store something in the address of a pointer.

Originally Posted by *DEAD*

ill use the type defs, but im still curious as to how to define a pointer.

as for the &, its technically the "correct" way to do it (so i believe), and in my opinion not using short hand makes code much more readable.

You can do either.

Short hand would be calling a variable "x" instead of "numElephantsInRoom." Leaving out an unnecessary ampersand isn't exactly conducting major surgery. I fail to see how its presence or absence in this case makes code any more/less clear.

I'd say 50% of the code I've seen uses it when taking function addresses, the other half doesn't.

EDIT: Also, your pointer type is something like:

Code:

int (*func)(int (*)(int));

Pointer to function which takes (pointer to function which takes int returning int) returning int.

However, for member function pointers in C++, you need to use &function. I think that's where the myth comes from. For ordinary function pointers, in both languages, function and &function are equivalent.

http://cboard.cprogramming.com/showthread.php?t=77504
http://www.newty.de/fpt/fpt.html#assign