How do I deal with many levels of indentation?

by

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.

More Related Contents:

  1. Why shouldn’t I mix tabs and spaces?
  2. What indentation is required for a case statement within a let statement?
  3. Why give me error? [closed]
  4. What is the difference between . (dot) and $ (dollar sign)?
  5. Good examples of Not a Functor/Functor/Applicative/Monad?
  6. Haskell ranges and floats
  7. Defining a function by equations with different number of arguments
  8. A Haskell function of type: IO String-> String
  9. Haskell: Parse error in pattern
  10. IO happens out of order when using getLine and putStr
  11. What is indexed monad?
  12. What does $ mean/do in Haskell?
  13. Why are polymorphic values not inferred in Haskell?
  14. Cabal not installing dependencies when needing profiling libraries?
  15. Library function to compose a function with itself n times
  16. Should do-notation be avoided in Haskell?
  17. Why doesn’t Haskell sequence these IO actions properly?
  18. Why doesn’t this function work if I use “[xs]” instead of “xs”?
  19. What is fusion in Haskell?
  20. Typeclass constraints on data declarations
  21. What does the `forall` keyword in Haskell/GHC do?
  22. How does deriving work in Haskell?
  23. Foldr/Foldl for free when Tree is implementing Foldable foldMap?
  24. What characters are permitted for Haskell operators?
  25. Converting IO Int to Int
  26. Convert String to Integer/Float in Haskell?
  27. Haskell – depth for each node in binary tree using Reader monad
  28. Why does OCaml sometimes require eta expansion?
  29. Explanation of “tying the knot”
  30. Return specific type within Haskell

Leave a Comment