Floating point primitive

[omnisci3nce]

Addresses #9

Overview

Threads the basic Float and Double primitives from libclang through the IR and then c/caml generation.

Features

• Adds a float into the doggo example
• Get the C value (it automatically is casted from double to float since OCaml doesnt have 32bit floats AFAIK) with Double_val
• If returning a struct back to OCaml we need it to be boxed and so the caml_copy_double function is used (just referenced the docs for this :) )

Future Improvements

I decided to just leave the fallback to val_int and int_val for now because if we want to do more intelligent matching there we probably need a way of either bubbling up errors or raising an exception and I don't want to touch that right now without having a proper discussion with you & others.

---

I'm going to test out type aliases next (I usually typedef float and double to f32/f64 respectively for example).