Haskell conditional error

Question

I have a problem with a Haskell code, I have the following:

takeMeHere cs m =
    |(find (==E) cs == Nothing && (checkNextStep pozxCrt+1 pozyCrt m) == True) = (Just E,[(pozxCrt+1,pozyCrt)] ++ m)
    |(find (==S) cs == Nothing && (checkNextStep pozxCrt pozyCrt-1 m) == True) = (Just S,[(pozxCrt,pozyCrt-1)] ++ m)
    |(find (==W) cs == Nothing && (checkNextStep pozxCrt-1 pozyCrt m) == True) = (Just W,[(pozxCrt-1,pozyCrt)] ++ m)
    |(find (==N) cs == Nothing && (checkNextStep pozxCrt pozyCrt+1 m) == True) = (Just N,[(pozxCrt,pozyCrt+1)] ++ m)
    |otherwise = (Nothing,m)
where
    pozxCrt=fst(head m)
    pozyCrt=snd(head m)

checkNextStep x y m = if(find (== (x,y)) m == Nothing) then True
   else False

I get a parse error on input "|" . If I write the code with something like if then else if then...it works. But I want to use the | for a more compact coding.What seems to be the problem here ?

Answer 1

To fix parsing error, remove = from first line. The = sign is put after the guards.

Next, you should indent "where"

takeMeHere cs m
    |(find (==E) cs == Nothing && (checkNextStep pozxCrt+1 pozyCrt m) == True) = (Just E,[(pozxCrt+1,pozyCrt)] ++ m)
    |(find (==S) cs == Nothing && (checkNextStep pozxCrt pozyCrt-1 m) == True) = (Just S,[(pozxCrt,pozyCrt-1)] ++ m)
    |(find (==W) cs == Nothing && (checkNextStep pozxCrt-1 pozyCrt m) == True) = (Just W,[(pozxCrt-1,pozyCrt)] ++ m)
    |(find (==N) cs == Nothing && (checkNextStep pozxCrt pozyCrt+1 m) == True) = (Just N,[(pozxCrt,pozyCrt+1)] ++ m)
    |otherwise = (Nothing,m)
  where
    pozxCrt=fst(head m)
    pozyCrt=snd(head m)

This will at least parse, yet it won't compile. The (checkNextStep pozxCrt pozyCrt+1 m) should be (checkNextStep pozxCrt (pozyCrt+1) m).

Let me add that you can fix a lot of verbose code:

• find (==E) cs == Nothing can be changed to E `notElem` x
• You do not need to compare with True: change x == True to x
• if x then True else False can be changed to x
• [x]++y can be changed to x:y
• You can use pattern matching like this: (pozxCrt, pozyCrt) = head m or (pozxCrt, pozyCrt):_ = m

The result is:

takeMeHere cs m                                                                 
    | E `notElem` cs && checkNextStep (pozxCrt+1) pozyCrt m = (Just E,(pozxCrt+1,pozyCrt):m)
    | S `notElem` cs && checkNextStep pozxCrt (pozyCrt-1) m = (Just S,(pozxCrt,pozyCrt-1):m)
    | W `notElem` cs && checkNextStep (pozxCrt-1) pozyCrt m = (Just W,(pozxCrt-1,pozyCrt):m)
    | N `notElem` cs && checkNextStep pozxCrt (pozyCrt+1) m = (Just N,(pozxCrt,pozyCrt+1):m)
    | otherwise = (Nothing,m)                                                   
  where                                                                         
    (pozxCrt, pozyCrt) = head m                                                 

checkNextStep x y m = (x,y) `notElem` m

You have a lot of repetition in the guards. A lot of repetition is a sign to create new functions.

move E (x, y) = (x+1, y) 
move S (x, y) = (x, y-1)
move N (x, y) = (x, y+1)
move W (x, y) = (x-1, y)

takeMeHere cs m
    | canGo E = go E
    | canGo S = go S
    | canGo W = go W
    | canGo N = go N
    | otherwise = (Nothing,m)
  where
    pos = head m
    canGo dir = dir `notElem` cs && checkNextStep (move dir pos) m
    go dir = (Just dir, move dir pos:m)

checkNextStep (x, y) m = (x,y) `notElem` m

Next step: use find canGo [E,S,W,N] to get rid of the guards:

 takeMeHere cs m =                                                               
    case find canGo [E,S,W,N] of                                                
      Just dir -> (Just dir, move dir pos:m)                                    
      Nothing -> (Nothing, m) 
    where ...