refactor chalk lowering

[nikomatsakis]

Chalk lowering grew rather organically and is kind of a mess. It should be separated out into distinct modules and phases. There was some recent discussion from the WG-traits gitter channel laying out some of the problems and ideas.

[scalexm]

Basically, we would have two new modules:

• a sub module ir/lowering which would deal with translating chalk-parse AST to chalk constructs (constructs found in ir/mod.rs)
• a new module rules which would deal with outputting program clauses and doing various things around it like default impls and WF requirements; this module could probably be split in multiple files, one for each type of rule

I guess we would keep the coherence module untouched, and we would probably continue to add things about specialization in that module?