Add an 'ignore' function to Control.Monad

[guest]

See  http://www.haskell.org/pipermail/libraries/2009-June/thread.html#11880

In short, add a 'ignore :: m a -> m ()' function to Control.Monad. This lets us do things like 'forkIO $ ignore stuff', as opposed to throwing around all sorts of '>> return ()'.

This function could be widely used by many libraries & apps, and has been repeatedly invented and suggested (see the thread). So far no one has said a word against it.

- -- | Convenience function. This is particularly good for 'forkIO' and 'forM_',
-- as they demand return types of 'IO ()', but most interesting IO functions
-- don't return void. So one can wrap them with 'ignore' (essentially a call to 'unit').
ignore :: Functor f => f a -> f ()
ignore = fmap (const ())

[guest]

 Nabble link for those who use it

[guest]

Patch adding ignore to base

[dons]

This is a generalization of the well-known FFI 'void' function:

module Foreign.Marshal.Error 

-- |Discard the return value of an 'IO' action
--
void     :: IO a -> IO ()
void act  = act >> return ()

in base for the last decade. Generalising 'void' (with that name), seems more intuitive to me.

[guest]

amended version; functor, name change, and more laconic documentation

[guest]

I've attached a new patch, one using the 'void' name and Functor type signature.

(I wondered about the section to put it in, but then I realized 'forever' wasn't in the Prelude either, so 'void' would be as happy there.)

[guest]

My final headcount of the very long discussion is here:  http://www.haskell.org/pipermail/libraries/2010-January/012972.html

I think it's clear that there's overwhelming support for adding void/ignore, with minor quibbling over the exact name & type signature. void :: f a -> f () is the best compromise, I think, and I hope now that Don's point about 'void' has been addressed, igloo or someone could add it to base.

[guest]

What needs to be done to get this applied? Since the patch/email/ticket flurry 4 days, there haven't even been any objections. Is there someone I need to specifically email or chat with in #haskell?

--gwern

[malcolm.wallace@…]

Patch applied. Should it be merged to the ghc-6.12.x branch?

[guest]

Malcolm: I don't see any reason why it shouldn't be, do you?

(I suppose this would involve issues of base's version-number, but I don't know anything about that.)

[ross]

We don't change the API of core packages during the cycle.