v1.2.0

This release focused mostly on performance. While performance is drastically
improved with this release, it won't be the end of things in a continous effort
to get neodoc as snappy as possible, while retaining it's strong parsing
capabilities and feature set.

Changes

• Update to purescript 0.10.x
• Improve overall performance #81
  • To give an idea, the uglify example now runs about 140ms faster than
    before!
  • Optimize large parts of the code-base (lexer, solver, arg-parser)
  • Pre-trim descriptions section to speed up descriptions lexing
  • Let go of purescript-parsing, use own parser everywhere
  • Avoid partial function application where feasible, especially if function
    takes many arguments.
  • Optimize many parser combinators to work especially well with chars
  • Run the compiled output through the closure compiler to bring down require
    times. This comes at the cost of readable output, but provides a rough
    20ms boost requiring neodoc, which is now around 50ms on the machines I
    tested on. Still not ideal, maybe we can shed more bloat.