I am writing a script that has a very logically complicated loop:
main = do
    inFH <- openFile "..." ReadMode
    outFH <- openFile "..." WriteMode
    forM myList $ \ item ->
        ...
        if ... 
            then ...
            else do
                ...
                case ... of
                    Nothing -> ...
                    Just x  -> do
                        ...
                            ...
The code soon flies to the right, so I was thinking breaking it into pieces, using for example where clauses. The problem is, many of these ... contain reading/writing statements to the two handles inFH and outFH, and using a where statement will render those two names out of context. I would have to send in these two variables everytime I use a where statement.
Is there a better way of dealing with this?
In many cases, these deeply-nested indentations are the result of deeply-nested error checking. If that's so for you, you should look into MaybeT and its big brother ExceptT. These offer a clean way to separate the "what do we do when something went wrong" code from the "what do we do assuming everything goes right" code. In your example, I might write:
data CustomError = IfCheckFailed | MaybeCheckFailed
main = handleErrors <=< runExceptT $ do
    inFH  <- liftIO $ openFile ...
    outFH <- liftIO $ openFile ...
    forM myList $ \item -> do
        when (...) (throwError IfCheckFailed)
        ...
        x <- liftMaybe MaybeCheckFailed ...
        ...
liftMaybe :: MonadError e m => e -> Maybe a -> m a
liftMaybe err = maybe (throwError err) return
handleErrors :: Either CustomError a -> IO a
handleErrors (Left err) = case err of
    IfCheckFailed    -> ...
    MaybeCheckFailed -> ...
handleErrors (Right success) = return success
Notice that we still increase indentation at the forM loop; but the other checks are done "in-line" in main, and are handled all at the same indentation level in handleErrors.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With