Add escape-to-raw MIR statement

Add a new MIR "ghost state statement": Escaping a ptr to permit raw accesses.

~~This includes #55549, [click here](https://github.com/RalfJung/rust/compare/miri-visitor...RalfJung:escape-to-raw) for just the new commits.~~

std: Synchronize access to global env during `exec`

This commit, after reverting #55359, applies a different fix for #46775
while also fixing #55775. The basic idea was to go back to pre-#55359
libstd, and then fix #46775 in a way that doesn't expose #55775.

The issue described in #46775 boils down to two problems:

* First, the global environment is reset during `exec` but, but if the
  `exec` call fails then the global environment was a dangling pointer
  into free'd memory as the block of memory was deallocated when
  `Command` is dropped. This is fixed in this commit by installing a
  `Drop` stack object which ensures that the `environ` pointer is
  preserved on a failing `exec`.

* Second, the global environment was accessed in an unsynchronized
  fashion during `exec`. This was fixed by ensuring that the
  Rust-specific environment lock is acquired for these system-level
  operations.

Thanks to Alex Gaynor for pioneering the solution here!

Closes #55775

This commit, after reverting #55359, applies a different fix for #46775
while also fixing #55775. The basic idea was to go back to pre-#55359
libstd, and then fix #46775 in a way that doesn't expose #55775.

The issue described in #46775 boils down to two problems:

* First, the global environment is reset during `exec` but, but if the
  `exec` call fails then the global environment was a dangling pointer
  into free'd memory as the block of memory was deallocated when
  `Command` is dropped. This is fixed in this commit by installing a
  `Drop` stack object which ensures that the `environ` pointer is
  preserved on a failing `exec`.

* Second, the global environment was accessed in an unsynchronized
  fashion during `exec`. This was fixed by ensuring that the
  Rust-specific environment lock is acquired for these system-level
  operations.

Thanks to Alex Gaynor for pioneering the solution here!

Closes #55775
Co-authored-by: NAlex Gaynor <alex.gaynor@gmail.com>

xLTO: Don't pass --plugin-opt=thin to LLD. That's not supported anymore.

It seems that `-plugin-opt=thin` is not needed anymore when invoking LLD for ThinLTO. Unfortunately, still passing the option makes LLD crash instead of giving a deprecation warning or something.

This reverts commit 36fe3b60.

Switch AppVeyor builds to the GCE cloud

[Recommended by AppVeyor][1] this isn't done by default for all builds
due to the high startup overhead (3-4 minutes for a VM), but that's
paltry compared to our overall build times so should be more than
applicable!

[1]: https://help.appveyor.com/discussions/questions/29832-did-recent-changes-apply-to-possibly-slow-down-builds#comment_46494058

chalk lowering rule: ProjectionEq-Normalize

cc #49177
r? @scalexm

Rollup of 20 pull requests

Successful merges:

 - #55136 (Remove short doc where it starts with a codeblock)
 - #55711 (Format BtreeMap::range_mut example)
 - #55722 (impl_stable_hash_for: support enums and tuple structs with generic parameters)
 - #55754 (Avoid converting bytes to UTF-8 strings to print, just pass bytes to stdout/err)
 - #55804 (rustdoc: don't inline `pub use some_crate` unless directly asked to)
 - #55805 (Move `static_assert!` into librustc_data_structures)
 - #55837 (Make PhantomData #[structural_match])
 - #55840 (Fix TLS errors when downloading stage0)
 - #55843 (add FromIterator<A> to Box<[A]>)
 - #55858 (Small fixes on code blocks in rustdoc)
 - #55863 (Fix a typo in std::panic)
 - #55870 (Fix typos.)
 - #55874 (string: Add documentation for `From` impls)
 - #55879 (save-analysis: Don't panic for macro-generated use globs)
 - #55882 (Reference count `crate_inherent_impls`s return value.)
 - #55888 (miri: for uniformity, also move memory_deallocated to AllocationExtra)
 - #55889 (global allocators: add a few comments)
 - #55896 (Document optimizations enabled by FusedIterator)
 - #55905 (Change `Lit::short_name` to `Lit::literal_name`.)
 - #55908 (Fix their/there grammar nit)

[Recommended by AppVeyor][1] this isn't done by default for all builds
due to the high startup overhead (3-4 minutes for a VM), but that's
paltry compared to our overall build times so should be more than
applicable!

[1]: https://help.appveyor.com/discussions/questions/29832-did-recent-changes-apply-to-possibly-slow-down-builds#comment_46494058

Check for negative impls when finding auto traits

Fixes #55321

When AutoTraitFinder begins examining a type, it checks for an explicit
negative impl. However, it wasn't checking for negative impls found when
calling 'select' on predicates found from nested obligations.

This commit makes AutoTraitFinder check for negative impls whenever it
makes a call to 'select'. If a negative impl is found, it immediately
bails out.

Normal users of SelectioContext don't need to worry about this, since
they stop as soon as an Unimplemented error is encountered. However, we
add predicates to our ParamEnv when we encounter this error, so we need
to handle negative impls specially (so that we don't try adding them to
our ParamEnv).

Use bound types in `librustc_traits`

r? @nikomatsakis
see the tests which now move the higher-ranked binders to the left in implied bound rules!

A `WF(Type)` predicate was used previously, which did not play
well with implied bounds in chalk.

impl_stable_hash_for: support enums and tuple structs with generic parameters

Port a bunch of implementations over to the macro, now that that is possible.

Remove short doc where it starts with a codeblock

Fixes #54975.

Fix their/there grammar nit

Change `Lit::short_name` to `Lit::literal_name`.

This avoids a moderately hot allocation in `parse_lit_token`.

r? @Mark-Simulacrum

Document optimizations enabled by FusedIterator

When reading this I wondered what “some significant optimizations” referred to. As far as I can tell from reading code, the specialization of `.fuse()` is the only case where `FusedIterator` has any impact at all. Is this accurate @Stebalien?

global allocators: add a few comments

These comments answer some questions that came up when I tried to understand how the control flow works for the global allocator, `Global` and `System`.

r? @alexcrichton

miri: for uniformity, also move memory_deallocated to AllocationExtra

r? @oli-obk

Reference count `crate_inherent_impls`s return value.

The repeated cloning of the result in `inherent_impls` queries has quite
an impact on crates with many inherent trait implementations.

For instance on https://github.com/jmesmon/stm32f429, `cargo check` went from 75 seconds to 38 seconds on my machine.

save-analysis: Don't panic for macro-generated use globs

Follow-up to https://github.com/rust-lang/rust/commit/c2bb7cadf24e82b80f403c09e800fe5fad504caf - as before, ignore the use globs in macro expansions.

Fixes https://github.com/rust-lang-nursery/rls/issues/1117.
Closes #55480.

r? @nrc

string: Add documentation for `From` impls

Hi this is part of #51430. I'm a first time contributor, so I started with a small task adding a bit of documentation for From impls.

Fix typos.

Fix a typo in std::panic

Small fixes on code blocks in rustdoc

It currently has a little margin issue with the `i` and there were js errors.

r? @QuietMisdreavus

add FromIterator<A> to Box<[A]>

Fix TLS errors when downloading stage0

While attempting to test #49878 on Windows I hit the following error when attempting to download stage0.

```
The request was aborted: Could not create SSL/TLS secure channel
```

Instead of using the shell, we can just use `urllib`, which seems to fix the issue.

Make PhantomData #[structural_match]

fixes https://github.com/rust-lang/rust/issues/55028

This makes `PhantomData<T>` structurally matchable, irrespective of whether `T` is, per the discussion on this week's language team meeting (the general consensus was that this was a bug-fix).

All types containing `PhantomData<T>` and which used `#[derive(PartialEq, Eq)]` and were previously not `#[structural_match]` only because of `PhantomData<T>` will now be `#[structural_match]`.

r? @nikomatsakis