# Difference between revisions of "Stateful nondeterminism"

From HaskellWiki

DonStewart (talk | contribs) (category) |
m |
||

Line 1: | Line 1: | ||

+ | From [[New monads]] |
||

+ | |||

+ | |||

If you want to do nondeterministic computation with local states for each of your threads and a global state shared by all your threads, use this monad: |
If you want to do nondeterministic computation with local states for each of your threads and a global state shared by all your threads, use this monad: |
||

<haskell> |
<haskell> |
||

Line 34: | Line 37: | ||

[[Category:Idioms]] |
[[Category:Idioms]] |
||

+ | [[Category:Code]] |

## Revision as of 08:47, 27 January 2007

From New monads

If you want to do nondeterministic computation with local states for each of your threads and a global state shared by all your threads, use this monad:

```
newtype SuperState s t a = SuperState { runSuperState :: (s -> t -> (t,[(a,s)])) }
instance Monad (SuperState s t) where
return a = SuperState $ \s t -> (t,[(a,s)])
(SuperState x) >>= f = SuperState $ \s t -> let (t',stateList) = x s t
in foldl (\(newt,sofar) (v,s) -> let (t'',lst) = runSuperState (f v) s newt
in
(t'',sofar++lst)) (t',[]) stateList
instance MonadPlus (SuperState s t) where
mzero = mz
mplus = mp
getGlobal = SuperState $ \s t-> (t,[(t,s)])
getLocal = SuperState $ \s t -> (t,[(s,s)])
putLocal s = SuperState $ \_ t -> (t,[((),s)])
putGlobal t = SuperState $ \s _ -> (t,[((),s)])
mz = SuperState $ \_ t -> (t,[])
mp (SuperState a) (SuperState b) =
SuperState $ \s t ->
let
(t',stateList) = a s t
(t'',stateList') = b s t'
in
(t'',stateList++stateList')
```