Question

I implemented a simple state machine in Python:

import time

def a():
    print "a()"
    return b

def b():
    print "b()"
    return c

def c():
    print "c()"
    return a


if __name__ == "__main__":
    state = a
    while True:
        state = state()
        time.sleep(1)

I wanted to port it to C, because it wasn’t fast enough. But C doesn’t let me make a function that returns a function of the same type. I tried making the function of this type: typedef *fn(fn)(), but it doesn’t work, so I had to use a structure instead. Now the code is very ugly!

#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>

typedef struct fn {
    struct fn (*f)(void);
} fn_t;

fn_t a(void);
fn_t b(void);
fn_t c(void);

fn_t a(void)
{
    fn_t f = {b};

    (void)printf("a()\n");

    return f;
}

fn_t b(void)
{
    fn_t f = {c};

    (void)printf("b()\n");

    return f;
}

fn_t c(void)
{
    fn_t f = {a};

    (void)printf("c()\n");

    return f;
}

int main(void)
{
    fn_t state = {a};

    for(;; (void)sleep(1)) state = state.f();

    return EXIT_SUCCESS;
}

So I figured it’s a problem with C’s broken type system. So I used a language with a real type system (Haskell), but the same problem happens. I can’t just do something like:

type Fn = IO Fn
a :: Fn
a = print "a()" >> return b
b :: Fn
b = print "b()" >> return c
c :: Fn
c = print "c()" >> return a

I get the error, Cycle in type synonym declarations.

So I have to make some wrapper the same way I did for the C code like this:

import Control.Monad
import System.Posix

data Fn = Fn (IO Fn)

a :: IO Fn
a = print "a()" >> return (Fn b)

b :: IO Fn
b = print "b()" >> return (Fn c)

c :: IO Fn
c = print "c()" >> return (Fn a)

run = foldM (\(Fn f) () -> sleep 1 >> f) (Fn a) (repeat ())

Why is it so hard to make a state machine in a statically typed language? I have to make unnecessary overhead in statically typed languages as well. Dynamically typed languages don’t have this problem. Is there an easier way to do it in a statically typed language?

Answer 1

If you use newtype instead of data, you don’t incur any overhead. Also, you can wrap each state’s function at the point of definition, so the expressions that use them don’t have to:

import Control.Monad

newtype State = State { runState :: IO State }

a :: State
a = State $ print "a()" >> return b

b :: State
b = State $ print "b()" >> return c

c :: State
c = State $ print "c()" >> return a

runMachine :: State -> IO ()
runMachine s = runMachine =<< runState s

main = runMachine a

Edit: it struck me that runMachine has a more general form; a monadic version of iterate:

iterateM :: Monad m => (a -> m a) -> a -> m [a]
iterateM f a = do { b <- f a
                  ; as <- iterateM f b
                  ; return (a:as)
                  }

main = iterateM runState a

Edit: Hmm, iterateM causes a space-leak. Maybe iterateM_ would be better.

iterateM_ :: Monad m => (a -> m a) -> a -> m ()
iterateM_ f a = f a >>= iterateM_ f

main = iterateM_ runState a

Edit: If you want to thread some state through the state machine, you can use the same definition for State, but change the state functions to:

a :: Int -> State
a i = State $ do{ print $ "a(" ++ show i ++ ")"
                ; return $ b (i+1)
                }

b :: Int -> State
b i = State $ do{ print $ "b(" ++ show i ++ ")"
                ; return $ c (i+1)
                }

c :: Int -> State
c i = State $ do{ print $ "c(" ++ show i ++ ")"
                ; return $ a (i+1)
                }

main = iterateM_ runState $ a 1