rustc_trans: clobber $1 (aka $at) on mips

This copies what clang does. There is a long explanation as to why this is needed in the clang source (tools/clang/lib/Basic/Targets/Mips.h).

std: use more portable error number in from_raw_os_error docs

On MIPS, error number 98 is not `EADDRINUSE` (it is `EPROTOTYPE`). To fix the resulting test failure this causes, use a more portable error number in the example documentation. `EINVAL` shold be more reliable because it was defined in the original Unix as 22 so hopefully most derivatives have defined it the same way.

Fix regression: account for trait methods in arg count mismatch error

Fixed #47706 (https://github.com/rust-lang/rust/issues/47706#issuecomment-361161495)

Original PR #47747 missed methods on trait definitions.

This edit was done in GitHub. I think I got the signature of the variant right, going by the ICE debug output and the other cases above.

Marked Debug implementations for primitive types as #[inline]

Change for issue #47792.

use correct casing for rename suggestions

If the original name is uppercase, use camel case. Otherwise, use snake
case.

Move Duration to libcore

Fixes #46520; should be merged after #46508.

This copies what clang does. There is a long explanation as to why this
is needed in the clang source (tools/clang/lib/Basic/Targets/Mips.h).

On MIPS, error number 98 is not EADDRINUSE (it is EPROTOTYPE). To fix the
resulting test failure this causes, use a more portable error number in
the example documentation. EINVAL shold be more reliable because it was
defined in the original Unix as 22 so hopefully most derivatives have
defined it the same way.

Rollup of 12 pull requests

- Successful merges: #47515, #47603, #47718, #47732, #47760, #47780, #47822, #47826, #47836, #47839, #47853, #47855
- Failed merges:

rustdoc: Fix link title rendering with hoedown

The link title needs to be HTML escaped.

It was broken by #47046.

r? @QuietMisdreavus

Increase test coverage of use_nested_groups

r? @nikomatsakis

Document that `Index` ops can panic on `HashMap` & `BTreeMap`.

Fixes https://github.com/rust-lang/rust/issues/47011.

Add per-stage RUSTFLAGS: RUSTFLAGS_STAGE_{0,1,2} and RUSTFLAGS_STAGE_NOT_0

Fixes #47658.

r? @alexcrichton

Whitelist v7 feature for ARM and AARCH64.

Needed for `v7` features in `coresimd`.

See https://github.com/rust-lang-nursery/stdsimd/blob/b2f7be24d5043a88427f9a5258ca9a51ede6d029/coresimd/src/arm/v7.rs#L40 which used to work but doesn't anymore.

r? alexcrichton

Whitelist aes x86 feature flag

Required to fix https://github.com/rust-lang-nursery/stdsimd/issues/295 in stdsimd.

Closes #44544 .

r? @alexcrichton

Add line numbers and columns to error messages spanning multiple files

If an error message is emitted that spans several files, only the
primary file currently has line and column data attached. This is
useful information, even in files other than the one in which the error
occurs. We can often work out which line and column the error
corresponds to in other files — in this case it is helpful to add them
(in the case of ambiguity, the first relevant line/column is picked,
which is still helpful than none).

implement Send for process::Command on unix

closes https://github.com/rust-lang/rust/issues/47751

Run rustfmt and add doc comments to libsyntax/ext/tt/macro_parser.rs

Similar to #47603

cc @theotherphil since you reviewed my other PR 😄

And because they are already assigned on #47603:

r? @arielb1

Make run-pass/env-home-dir.rs test more robust

Remove the assumption that home_dir always returns Some.

This allows the test to be executed with [cross](https://github.com/japaric/cross).

Run rustfmt and add doc comments to libsyntax/ext/tt/quoted.rs

I was already going to try to understand this code to try to implement rust-lang/rfcs#2298. I figured I would put that knowledge into comments and share... This PR contains _no changes_ to the code itself -- just formatting and comments.

I'm not sure what the best way to do this is. I plan to make more such PR for other files, but I figured it would have fewer conflicts if I do it file by file...

Punctuation and clarity fixes.

syntax: Lower priority of `+` in `impl Trait`/`dyn Trait`

Now you have to write `Fn() -> (impl A + B)` instead of `Fn() -> impl A + B`, this is consistent with priority of `+` in trait objects (`Fn() -> A + B` means `(Fn() -> A) + B`).

To make this viable I changed the syntax to also permit `+` in return types in function declarations
```
fn f() -> dyn A + B { ... } // OK, don't have to write `-> (dyn A + B)`

// This is acceptable, because `dyn A + B` here is an isolated type and
// not part of a larger type with various operator priorities in play
// like `dyn A + B` in `Fn() -> dyn A + B` despite syntax similarities.
```
but you still have to use `-> (dyn A + B)` in function types and function-like trait object types (see this PR's tests for examples).

This can be a breaking change for code using `impl Trait` on nightly. The thing that is most likely to break is `&impl A + B`, it needs to be rewritten as `&(impl A + B)`.

cc https://github.com/rust-lang/rust/issues/34511 https://github.com/rust-lang/rust/issues/44662 https://github.com/rust-lang/rfcs/pull/438

Replace "lvalue" terminology with "place".

See #46425 for the previous PR (which only changed MIR-related code).

r? @nikomatsakis

The link title needs to be HTML escaped.

Remove the assumption that home_dir always returns Some

This allows the test to be executed with
[cross](https://github.com/japaric/cross).

Fixes #47658.

Make region inference use a dirty list

r? @nikomatsakis

rustc: Split Emscripten to a separate codegen backend

This commit introduces a separately compiled backend for Emscripten, avoiding
compiling the `JSBackend` target in the main LLVM codegen backend. This builds
on the foundation provided by #47671 to create a new codegen backend dedicated
solely to Emscripten, removing the `JSBackend` of the main codegen backend in
the process.

A new field was added to each target for this commit which specifies the backend
to use for translation, the default being `llvm` which is the main backend that
we use. The Emscripten targets specify an `emscripten` backend instead of the
main `llvm` one.

There's a whole bunch of consequences of this change, but I'll try to enumerate
them here:

* A *second* LLVM submodule was added in this commit. The main LLVM submodule
  will soon start to drift from the Emscripten submodule, but currently they're
  both at the same revision.
* Logic was added to rustbuild to *not* build the Emscripten backend by default.
  This is gated behind a `--enable-emscripten` flag to the configure script. By
  default users should neither check out the emscripten submodule nor compile
  it.
* The `init_repo.sh` script was updated to fetch the Emscripten submodule from
  GitHub the same way we do the main LLVM submodule (a tarball fetch).
* The Emscripten backend, turned off by default, is still turned on for a number
  of targets on CI. We'll only be shipping an Emscripten backend with Tier 1
  platforms, though. All cross-compiled platforms will not be receiving an
  Emscripten backend yet.

This commit means that when you download the `rustc` package in Rustup for Tier
1 platforms you'll be receiving two trans backends, one for Emscripten and one
that's the general LLVM backend. If you never compile for Emscripten you'll
never use the Emscripten backend, so we may update this one day to only download
the Emscripten backend when you add the Emscripten target. For now though it's
just an extra 10MB gzip'd.

Closes #46819

Fixes https://github.com/rust-lang/rust/issues/47011.

This commit introduces a separately compiled backend for Emscripten, avoiding
compiling the `JSBackend` target in the main LLVM codegen backend. This builds
on the foundation provided by #47671 to create a new codegen backend dedicated
solely to Emscripten, removing the `JSBackend` of the main codegen backend in
the process.

A new field was added to each target for this commit which specifies the backend
to use for translation, the default being `llvm` which is the main backend that
we use. The Emscripten targets specify an `emscripten` backend instead of the
main `llvm` one.

There's a whole bunch of consequences of this change, but I'll try to enumerate
them here:

* A *second* LLVM submodule was added in this commit. The main LLVM submodule
  will soon start to drift from the Emscripten submodule, but currently they're
  both at the same revision.
* Logic was added to rustbuild to *not* build the Emscripten backend by default.
  This is gated behind a `--enable-emscripten` flag to the configure script. By
  default users should neither check out the emscripten submodule nor compile
  it.
* The `init_repo.sh` script was updated to fetch the Emscripten submodule from
  GitHub the same way we do the main LLVM submodule (a tarball fetch).
* The Emscripten backend, turned off by default, is still turned on for a number
  of targets on CI. We'll only be shipping an Emscripten backend with Tier 1
  platforms, though. All cross-compiled platforms will not be receiving an
  Emscripten backend yet.

This commit means that when you download the `rustc` package in Rustup for Tier
1 platforms you'll be receiving two trans backends, one for Emscripten and one
that's the general LLVM backend. If you never compile for Emscripten you'll
never use the Emscripten backend, so we may update this one day to only download
the Emscripten backend when you add the Emscripten target. For now though it's
just an extra 10MB gzip'd.

Closes #46819