Document HIR structure

[rylev]

As a follow-up to #17, we still don't seem to document the structure of the HIR. I assume this would entail examining a bit of what specifically makes the HIR different from the AST.

[camelid]

I assume this would entail examining a bit of what specifically makes the HIR different from the AST.

One of the main differences is that it's desugared. So, for example, the AST has ExprKind::Try for expr?, but the HIR does not.

Another difference is that IIUC macros are fully expanded by the time the HIR is constructed.