First class support for Liquid Haskell?

Should we add first class support for Liquid Haskell in stack? I have not used Liquid Haskell much yet, so I don't know how much sense this makes. It could be as simple as `stack liquid`, and it will automatically install it appropriately if necessary, and run the tool on your code. I will send an email to the Liquid Haskell authors and get them involved in this discussion.

It is quite possible that first class support is unnecessary. I just found this project - https://github.com/spinda/liquidhaskell-cabal/blob/0.1.1.0/README.md - but I have not yet tried it, sounds like it supports stack decently. Perhaps it would be helpful to streamline things for usage with stack?

In https://github.com/fpco/store/pull/71 , @kantp added use of liquid haskell to verify that we stay within the bounds of the bytebuffer! This allows us to have code that is both fast and trustworthy. Fantastic!

Considering we are now maintaining code that uses Liquid Haskell, it only makes sense to have our Travis CI for `store` use it. To me, the most sensible implementation approach is by having direct support in stack.