Rollup of 8 pull requests

Successful merges:

 - #50531 (Cleanup uses of TypeIdHasher and replace them with StableHasher)
 - #50819 (Fix potential divide by zero)
 - #50827 (Update LLVM to 56c931901cfb85cd6f7ed44c7d7520a8de1edf97)
 - #50829 (CheckLoopVisitor: also visit break expressions)
 - #50854 (in which the unused shorthand field pattern debacle/saga continues)
 - #50858 (Reorder description for snippets in rustdoc documentation)
 - #50883 (Fix warning when building stage0 libcore)
 - #50889 (Update clippy)

Failed merges:

Update clippy


# Conflicts:
#	src/Cargo.lock

Fix potential divide by zero

This should fix #50761

I had trouble reproducing with the provided code, but looking at the stack trace would indicate that this code is the likely cause. I made a number of assumptions here, because I don't have enough context on how the register size is set:

1. I assumed `rest.unit.size.bytes()` can be 0, and it's ok if it's set to 0 before this function is called
2. I assumed that if `rest.unit.size.bytes()` is 0, that we want `rest_count` to also be 0.

Fix warning when building stage0 libcore

When building stage0 a warning will be triggered when compiling libcore
due to `align_to_offsets` not being used.

Reorder description for snippets in rustdoc documentation

The example code snippets for the `no_run` and `compile_fail` attributes in the rustdoc documentation were followed by the description for the wrong attribute.  This patch reorders the descriptions to match the code snippets.

Rollup merge of #50854 - zackmdavis:and_the_case_of_the_unused_field_pattern_3_straight_to_video, r=estebank

in which the unused shorthand field pattern debacle/saga continues

In e4b1a797 (#47922), we corrected erroneous suggestions for unused
shorthand field pattern bindings, suggesting `field: _` where the
previous suggestion of `_field` wouldn't even have compiled
(#47390). Soon, it was revealed that this was insufficient (#50303), and
the fix was extended to references, slices, &c. (#50327) But even this
proved inadequate, as the erroneous suggestions were still being issued
for patterns in local (`let`) bindings (#50804). Here, we yank the
shorthand-detection and variable/node registration code into a new
common function that can be called while visiting both match arms and
`let` bindings.

Resolves #50804.

r? @estebank

CheckLoopVisitor: also visit break expressions

Fixes #50802

Update LLVM to 56c931901cfb85cd6f7ed44c7d7520a8de1edf97

This brings in https://github.com/rust-lang/llvm/pull/115, which fixes https://github.com/rust-lang/rust/issues/49873.

Prevent main from having a where clause.

Closes #50714

Should this have a crater run?

cc #48557, #48214

r? @nikomatsakis

Ensure that statics are always ByRef

Statics aren't values to be used, they are names for memory locations.

r? @EddyB

cc @Zoxc

fixes #50706

remove semicolon -_-

Add rem_bytes to conditional to avoid error when performing mod by 0

Add test file to confirm compilation passes.

Ensure we don't divide or mod by zero in llvm_type. Include test file from issue.

Cleanup uses of TypeIdHasher and replace them with StableHasher

Fixes #50424

r? @michaelwoerister

Add lint checks for unused loop labels

Previously no warning was generated when a loop label was written out but never used (i.e. in a `break` or `continue` statement):
```rust
fn main() {
    'unused_label: loop {}
}
```
would compile without complaint.

This fix introduces `-W unused_loop_label`, which generates the following warning message for the above snippet:
```
warning: unused loop label
 --> main.rs:2:5
  |
2 |     'unused_label: loop {}
  |     ^^^^^^^^^^^^^
  |
  = note: #[warn(unused_loop_label)] on by default
```

Fixes: #50751.

Turn deprecation lint `legacy_imports` into a hard error

Closes https://github.com/rust-lang/rust/issues/38260

The lint was introduced in Dec 2016, then made deny-by-default in Jun 2017 when crater run found 0 regressions caused by it.

This lint requires some not entirely trivial amount of import resolution logic that (surprisingly or not) interacts with `feature(use_extern_macros)` (https://github.com/rust-lang/rust/issues/35896), so it would be desirable to remove it before stabilizing `use_extern_macros`.
In particular, this PR fixes the failing example in https://github.com/rust-lang/rust/issues/50725 (but not the whole issue, `use std::panic::{self}` still can cause other undesirable errors when `use_extern_macros` is enabled).

use `reset_unifications` instead of creating new unification table

This eliminates a hot spot in NLL performance.

Emit noalias on &mut parameters by default

This used to be disabled due to LLVM bugs in the handling of
noalias information in conjunction with unwinding. However,
according to #31681 all known LLVM bugs have been fixed by
LLVM 6.0, so it's probably time to reenable this optimization.

-Z no-mutable-noalias is left as an escape-hatch to debug problems
suspected to stem from this change.

Revert #50105 until regression is fixed

Discovered at https://github.com/rust-lang/rust/pull/50105#issuecomment-388630750 it looks like this caused a regression with i686 musl, so let's revert in the meantime while a fix is worked out

When building stage0 a warning will be triggered when compiling libcore
due to align_to_offsets not being used.

 rustc_mir: allow promotion of promotable temps indexed at runtime.

Fixes #49955.

r? @nikomatsakis

This reverts commit b2a2450cf2925498e1c3bdd781aa1978f9eb00e5.

This reverts commit b9257e2ca161b1bf5aae9d6b667f4d0c6b8d7be6.

This functionality is being reimplemented in the resolver phase

This reverts commit 503a69e844970476b27bf1ac7be951bb22194f50.