You could perhaps write MyWierdMonadT and MyAnotherMonadT.
Also, probably layers package [1] is related to what you ask.
But if you actually want arbitrary nested monads like m1 (m2 (m3 (m4 a))),
there's no simple way.
Consider Maybe [MyType Int]. MyType may have or may not have a
Monadinstance. Even more: it can have instance in one module, but not
have in
another. But suppose MyType has a Num instance. Then how this should
work: return
2 :: Maybe [MyType Int]? Should 2 be of type Int or MyType Int?
Note that you can make an instance Num a => Num [a] which makes sense
(lists as polynoms). So the same question arises even for Maybe [Int].
If you **really want to do it** (which I doubt), you can use TH (do
determine how many return's to use, though it's ambiguous anyway).
[1]
http://hackage.haskell.org/package/layers-0.1/docs/Documentation-Layers-Over...
2013/11/23 Dmitry Bogatov
Perhaps you just want monad transformers [1, 2]. If you're not familiar with them, you should probably read [3].
With transformers you'd be able to do this (and more):
$ return 2 :: IO Int $ return 2 :: ListT Maybe Int -- this works like Maybe [Int] $ return 2 :: MaybeT [] Int -- this works like [Maybe Int]
In fact, I read and used transformers a bit. If I understand correctly, I can only do such things for monads, supported by transformers. So,
I cant get (return 2) :: MyWierdMonad (MyAnotherMonad 2)
Yes, I know, it have little practical use, but I am insterested about teoretical part of question, magic of ghc.
-- Best regards, Dmitry Bogatov
, Free Software supporter and netiquette guardian. git clone git://kaction.name/rc-files.git --depth 1 GPG: 54B7F00D Html mail and proprietary format attachments are forwarded to /dev/null.