refactor to remove trans::adt and make rustc::ty::layout authoritative

I asked on IRC about optimizing struct layout by reordering fields from most-aligned to least-aligned and somehow ended up getting talked into doing this.  The goal here is to make `layout` authoritative and to remove `adt`.  The former has been accomplished by reimplementing `represent_type_uncached` and the latter is in progress.  @EddyB thought I should make the PR now.

My plan is to reserve the actual optimization for a second PR, as this work is useful by itself.

rustbuild: Nicer error for host builds of targets

If a triple is configured only as a target, not a host, then trying to build
that triple with host artifacts would cause a panic. Fail a little nicer
instead.

Closes #36268

Load macros from `#[macro_use]` extern crates in `resolve`

r? @nrc

Docs: Update to "Ownership" section

Found #34865 while looking for easy docs wins I could contribute with.

don't unwrap non-local span

Fixes #36708.

Skip binary tidy check when on Windows Linux Subsystem

While it's possible that other linux systems will include "Microsoft" in
their /proc/version, this is deemed unlikely, and since this is a tidy
check, will likely be caught by buildbot/travis either way.

Fixes #36706.

While it's possible that other linux systems will include "Microsoft" in
their /proc/version, this is deemed unlikely, and since this is a tidy
check, will likely be caught by buildbot/travis either way.

[breaking-change] Remove TypedArena::with_capacity

This is a follow-up to #36592.

The function is unused by rustc. Also, it doesn't really follow the
usual meaning of a `with_capacity` function because the first chunk
allocation is now delayed until the first `alloc` call.

This change reduces the size of `TypedArena` by one `usize`.

@EddyB: we discussed this on IRC. Would you like to review it?

librustc_mir: Propagate constants during copy propagation.

This optimization kicks in a lot when bootstrapping the compiler.

r? @EddyB

This optimization kicks in a lot when bootstrapping the compiler.

Revert "implement `From<Vec<char>>` and `From<&'a [char]>` for `String`"

This reverts commit ac73335f.

This is a revert of https://github.com/rust-lang/rust/pull/35054, which resulted in at least 7 known regressions, reported [here](https://internals.rust-lang.org/t/regression-report-stable-2016-08-16-vs-beta-2016-09-21/4119) and [here](https://github.com/rust-lang/rust/issues/36352), which will hit stable next week.

I think this breakage was somewhat unanticipated, and we did not realize so many crates were broken until this week, so reverting is the conservative thing to do until we figure out how not to cause so much breakage. I've run crater on the revert and did not find any new breakage from the revert.

Fixes https://github.com/rust-lang/rust/issues/36352

cc @pwoolcoc @rust-lang/libs

ICH: Add ability to test the ICH of exported metadata items.

Also adds an example test case for ICH testing.

r? @nikomatsakis

Rollup of 7 pull requests

- Successful merges: #36018, #36498, #36500, #36559, #36566, #36578, #36664
- Failed merges:

Minor `VecDeque` doc examples cleanup.

None

Replace 'e.g.' by 'i.e.'

Fixes #36577.

r? @steveklabnik

Add regression test for #9837.

Fixes https://github.com/rust-lang/rust/issues/9837

Fix a variety of minor issues

CSS:
- use integer values for font-size in CSS
- use correct ordering of @import
- "invisible" isn't a tag - presume its a class
- "border-color" defines the complete border

python:
- use "not" instead of == "[]" for python
- prefer triple quoted docstrings
- prefer static functions where possible
- prefer modern style classes where possible
- remove semicolons;

global:
- remove duplicated words words

This installs documentation not man pages

This fixes the description for docdir in configure, it was my mistake for leaving it as "man pages", oops.

Fix wording for out-of-crate macro error

This fixes the wording of the note for out-of-crate macro errors to fix https://github.com/rust-lang/rust/issues/36469

The previous wording came from older logic in the PR that was replaced without updating the note.

strengthen doc warning about CString::from_raw

Saw unsound code using this function on IRC.

Load extern crates in `resolve`

This PR loads `extern crate`s in `resolve`'s `BuildReducedGraphVistor`.

r? @nikomatsakis

This reverts commit ac73335f.

Don't let a type parameter named "Self" unchanged past HIR lowering.

Fixes #36638 by rewriting `Self` type parameters (which are a parse error) to a `gensym("Self")`.

Background: #35605 introduced code across rustc that determines `Self` by its keyword name.
Reverting the sanity checks around that would inadvertently cause confusion between the true `Self` of a `trait` and other type parameters named `Self` (which have caused parse errors already).

I do not like to use `gensym`, and we may do something different here in the future, but this should work.

Fix ICE test in compiletest fail-tests

While working on Clippy which uses *compiletest*, I noticed that as long as all expected error are found, *compile-fail* tests will be marked *ok* even if there is an ICE. One function seems to have not been updated with JSON errors because ICEs are now reported like this:
```json
{"message":"../src/librustc/ty/context.rs:161: Attempted to intern `_` which contains inference types/regions in the global type context","code":null,"level":"error: internal compiler error","spans":[],"children":[],"rendered":null}
```
I don't think I can add a test for that.

I guess:
r? @nikomatsakis

Implement Debug for DirEntry.

None

resolve: groundwork for building the module graph during expansion

r? @nrc

Adds a `ProcMacro` form of syntax extension

This commit adds syntax extension forms matching the types for procedural macros 2.0 (RFC #1566), these still require the usual syntax extension boiler plate, but this is a first step towards proper implementation and should be useful for macros 1.1 stuff too.

Supports both attribute-like and function-like macros.

Note that RFC #1566 has not been accepted yet, but I think there is consensus that we want to head in vaguely that direction and so this PR will be useful in any case. It is also fairly easy to undo and does not break any existing programs.

This is related to #35957 in that I hope it can be used in the implementation of macros 1.1, however, there is no direct overlap and is more of a complement than a competing proposal. There is still a fair bit of work to do before the two can be combined.

r? @jseyfried

cc @alexcrichton, @cgswords, @EddyB, @aturon