Out-of-date state using state monad

Question

Here's a function runGame using the state monad:

runGame = do
    state <- get
    addPoints
    let newState = ...modify state in some way...
    put newState

1. I get the state
2. I modify it
3. I put the new state.

Now suppose I modify the state in addPoints too:

addPoints = do
  modify (+1)

The problem is, the state I have in runGame is out of date now. So when I modify that state and put it back, I've nullified the changes from addPoints .

What's a good way to solve this? I dont want to re- get the state after every function call "just in case" that function modified the state. I can switch to always using modify instead of get and put but that makes my code awkward.

Answer 1

You're looking for atomic/transactional modifications to the state—that's pretty much exactly what modify provides. It's not possible to "slip" some state modifying operation between the get and the put with modify . If you don't use it you'll have to manage the complexity of ensuring that nothing happens between get and put .