Implement RFC 2523, `#[cfg(version(..))]`

Hi! This is my first contribution to rust, I hope I didn't miss anything. I tried to implement this feature so that `#[cfg(version(1.44.0))]` works but the parser was printing an error that I wasn't sure how to fix so I just opted for implementing `#[cfg(version("1.44.0"))]` (note the quotes).

Tracking issue: #64796

`slice::fill`: use `T` instead of generic arg

implements https://github.com/rust-lang/rust/issues/70758#issuecomment-613994427

As the discussion in #70758 has shifted, I now use `T` instead of `&T`.

Don't bust caches on x.py check/build switches

Fixes #71152

Rollup of 7 pull requests

Successful merges:

 - #69274 (Implement RFC 2396: `#[target_feature]` 1.1)
 - #71767 (doc: make Stack and StackElement a little pretty)
 - #71772 (Mark query function as must_use.)
 - #71777 (cleanup: `config::CrateType` -> `CrateType`)
 - #71784 (Remove recommendation for unmaintained dirs crate)
 - #71785 (Update comment regarding SO_REUSEADDR on Windows)
 - #71787 (fix rustdoc warnings)

Failed merges:

r? @ghost

The extracted function can be used by the rest of bootstrap to detect if we've
already built an up-to-date LLVM (and so it's safe for us to either request it
or pretend it exists).

Implement RFC 2396: `#[target_feature]` 1.1

Tracking issue: #69098

r? @nikomatsakis
cc @gnzlbg @joshtriplett

fix rustdoc warnings

Update comment regarding SO_REUSEADDR on Windows

Remove recommendation for unmaintained dirs crate

See #71684 for reasoning here

cleanup: `config::CrateType` -> `CrateType`

Mark query function as must_use.

And use the `ensure()` version when the result is not needed.

doc: make Stack and StackElement a little pretty

Also, fix rustdoc warnings.

Implement RFC 2396: `#[target_feature]` 1.1

Tracking issue: #69098

r? @nikomatsakis
cc @gnzlbg @joshtriplett

update Miri

This contains the concurrency support by @vakaras :)

Fixes https://github.com/rust-lang/rust/issues/71729
r? @ghost
Cc @rust-lang/miri

Make clippy a git subtree instead of a git submodule

r? @EddyB

cc #70651

documentation at https://github.com/rust-lang/rust/pull/70654

Rollup of 6 pull requests

Successful merges:

 - #71712 (Miri: port error backtraces to std::backtrace)
 - #71736 (bootstrap: also apply unused-attributes hack without deny_warnings)
 - #71738 (remove AllocId generalization of Pointer)
 - #71739 (remove obsolete comment)
 - #71781 (Uncomment test code for failure to use `Box::pin`)
 - #71782 (Use a non-existent test path instead of clobbering /dev/null)

Failed merges:

r? @ghost

Use a non-existent test path instead of clobbering /dev/null

Fixes #71502.
r? @Mark-Simulacrum

Uncomment test code for failure to use `Box::pin`

Close #69083.

remove obsolete comment

Referenced was removed in 9f492fef

remove AllocId generalization of Pointer

This was only needed for the "snapshot" machinery, which is gone.

r? @oli-obk

bootstrap: also apply unused-attributes hack without deny_warnings

This is a follow-up to https://github.com/rust-lang/rust/pull/70881 that also silences these warnings when deny_warnings is off. They otherwise spam my screen during development and make it hard to see actual warnings.

Cc  @EddyB r? @Mark-Simulacrum

Miri: port error backtraces to std::backtrace

No need to pull in an external dependency if libstd already includes this feature (using the same dependency internally, but... still).

r? @oli-obk

git-subtree-dir: src/tools/clippy
git-subtree-mainline: 06c44816
git-subtree-split: d2708873

Co-authored-by: NJonas Schievink <jonasschievink@gmail.com>

wf: handle "livelock" checking before reaching `WfPredicates::compute`.

For `wf::obligations`'s "livelock" handling, this PR shouldn't cause any behavioral changes, as the check moved to it should be equivalent to the old one in `WfPredicates::compute`.

However, it fixes #70168 by making *other* users of `WfPredicates::compute` (that is, `wf::predicate_obligations` and `compute`'s own upvar handling) correct for `ty::Infer`, in that they now get a `WellFormed(ty::Infer(_))` obligation instead of silently ignoring the type.

r? @nikomatsakis

Rename `bitcode-in-rlib` option to `embed-bitcode`

This commit finishes work first pioneered in #70458 and started in #71528.
The `-C bitcode-in-rlib` option, which has not yet reached stable, is
renamed to `-C embed-bitcode` since that more accurately reflects what
it does now anyway. Various tests and such are updated along the way as
well.

This'll also need to be backported to the beta channel to ensure we
don't accidentally stabilize `-Cbitcode-in-rlib` as well.