Changes:

- Refactor move mode computation
- Removes move mode arguments, unary move, capture clauses
  (though they still parse for backwards compatibility)
- Simplify how moves are handled in trans
- Fix a number of illegal copies that cropped up
- Workaround for bug involving def-ids in params
  (see details below)

Future work (I'll open bugs for these...):

- Improve error messages for moves that are due
  to bindings
- Add support for moving owned content like a.b.c
  to borrow check, test in trans (but I think it'll
  "just work")
- Proper fix for def-ids in params

Def ids in params:

Move captures into a map instead of recomputing.

This is a workaround for a larger bug having to do with the def-ids associated
with ty_params, which are not always properly preserved when inlining.  I am
not sure of my preferred fix for the larger bug yet.  This current fix removes
the only code in trans that I know of which relies on ty_param def-ids, but
feels fragile.

Add a new method_super origin for supertrait methods. Also make
coherence create a table that maps pairs of trait IDs and self types
to impl IDs, so that it's possible to check a supertrait method
knowing only its index in its trait's methods (without knowing all
supertraits for a given trait).

r=nmatsakis and graydon -- with hope, we'll revamp all of this code as
per #4678, but for now this fixes the bug.

Closes #3979

"Dual impls" are impls that are both type implementations and trait
implementations. They can lead to ambiguity and so this patch removes them
from the language.

This also enforces coherence rules. Without this patch, records can implement
traits not defined in the current crate. This patch fixes this, and updates
all of rustc to adhere to the new enforcement. Most of this patch is fixing
rustc to obey the coherence rules, which involves converting a bunch of records
to structs.

r=brson

The tail portion of the pattern effectively borrows a vector,
but the borrow checker knew nothing about this.

r=catamorphism

contents of other borrowed pointers to the lifetimes of the
borrowed value.  Fixes #3148.

r=catamorphism

specifically:

freevars::freevar_entry
ty::{field_ty,AutoAdjustment,AutoRef}
mod::{method_param,method_map_entry}

r=pcwalton

Previously, the typechecker suppressed many but not all errors,
by suppressing errors where the actual type was either ty_err, or
a function type whose result was ty_err. Added a has_ty_err flag
to the type flags so as to suppress errors for any types involving
ty_err.

r=nmatsakis

This is an interim fix to address the "Beware!" unsoundness.  I have a
more comprehensive rewrite of mode.rs in the pipeline.

r=pcwalton

- Make `extern fn()` assignable to any closure type, rather than
  a subtype.
- Remove unused int_ty_set and float_ty_set
- Refactor variable unification and make it more DRY
- Do fn sub/lub/glb on the level of fn_sig
- Rename infer::to_str::ToStr to infer::to_str::InferStr
- Capitalize names of various types
- Correct hashing of FnMeta
- Convert various records-of-fns into structs-of-fns.  This is both
  eliminating use of deprecated features and more forwards compatible
  with fn reform.

r=pcwalton

Revert "librustc: Make unqualified identifier searches terminate at the nearest module scope. r=tjc"

This reverts commit a8d37af2.

r=brson

cc #3681

~20% perf win for trans on -O0, with other minor improvements across the board.
No effect on -O2.

librustc: Fix unconditional ty_to_str call in ty.rs. Was heaviest path in allocation. Negligible perf win. rs=perf-fix

librustc: Terminate name searches at the nearest module scope for paths that contain at least two components. r=graydon

For example, in let x: char = 42; This was an ICE and is now a
proper type error, as per #3477

librustc: Have coherence check Copy kind bounds when determining whether parameter substitutions could possibly unify. r=nmatsakis