Multiple errors during name resolution

[paf31]

@garyb @joneshf Could you please review? Thanks.

[paf31]

Resolves #524.

[coveralls]

Coverage remained the same when pulling 9eb1965 on 524 into 33db511 on master.

[garyb]

Oh, didn't realise it was this easy! 👍

[paf31]

Yeah :) parU is supposed to just be a drop-in replacement for mapM but ideally, we would use an Applicative wrapper I guess.

Anyway, I'll have a stab at Names and then this should be done, yes?

[garyb]

Yep yep, sounds good 🍰

[paf31]

Ok, I think this is good to go. How does it look?

[garyb]

Does TypeDeclarations need parU too?

[paf31]

I left it out for now since desugarTypeDeclarations isn't written using mapM, although maybe it should be.

[garyb]

Ah ok, everything else looks good to me anyway. 👍

[paf31]

Actually, this is exactly where using Applicative in a newtype would have been a better approach.