Applicative-Monad Proposal fixes.

[barrucadu]

All Monads defined in the code are now Functors and Applicatives, to comply with the applicative-monad proposal.

All praise our Glorious Glasgow overlords!

[barrucadu]

I agree with the points in that answer, monad transformers are good if they simplify things. For instance, in yukibot I use transformers to avoid explicitly passing around a configuration parameter everywhere. Whilst I haven't looked at the F-lite code much, the monads there all look fairly simple and very different, so I'm not sure transformers would give any benefit.

Regarding 7.10, yes, it works, there's just a compiler warning about the import being redundant. There are, however, other things which don't work in 7.10:

Flite/ParseLib.hs:13:1: Warning:
    Local definition of ‘pure’ clashes with a future Prelude name - this will become an error in GHC 7.10, under the Applicative-Monad Proposal.

Flite/ParseLib.hs:16:3: Warning:
    Local definition of ‘<*>’ clashes with a future Prelude name - this will become an error in GHC 7.10, under the Applicative-Monad Proposal.

Flite/Dependency.hs:33:1: Warning:
    Local definition of ‘join’ clashes with a future Prelude name - this will become an error in GHC 7.10, under the Applicative-Monad Proposal.

Flite/CallGraph.hs:26:1: Warning:
    Local definition of ‘join’ clashes with a future Prelude name - this will become an error in GHC 7.10, under the Applicative-Monad Proposal.

If ParseLib gets dropped (as mused over in #1), the first two are a non-issue. For the latter two, I'm not sure if it would be better to just shadow the Prelude join, or to change the names of the functions.