How does XMonad keys functions gets reduced?

Question

I'm trying to understand how XMonad configuration functions are evaluated. The first example is from Xmonad docs:

import qualified Data.Map as M
--

main = xmonad $ defaultConfig {
  --
  , keys          = \c -> mykeys c `M.union` keys defaultConfig c
  --
  }
where
    mykeys (XConfig {modMask = modm}) = M.fromList $
         [ ((modm , xK_x), spawn "xlock") ]

The type of keys is a function: keys :: !(XConfig Layout -> Map (ButtonMask, KeySym) (X ())) The question here is evaluation order. keys returns Map when applied to some XConfig, and then Map we've got should be applied to c (if I get it right), but how Map can be applied to something?

Answer 1

Your description isn't quite correct. keys is a record field.

When you have a record type R with a field foo :: T, then foo itself (used as a value) has the type R -> T (it's a record accessor function).

Since keys is a field of XConfig, its type is actually

keys :: XConfig l -> !(XConfig Layout -> Map (ButtonMask, KeySym) (X ()))

In

keys defaultConfig c

We extract the keys field from defaultConfig. The value stored in this field is itself a function, which we apply to c to get a Map.