Co-authored-by: Tomoaki Kawada <[email protected]>

When a binding is declared without a value, borrowck verifies that all
codepaths have *one* assignment to them to initialize them fully. If
there are any cases where a condition can be met that leaves the binding
uninitialized or we attempt to initialize a field of an unitialized
binding, we emit E0381.

We now look at all the statements that initialize the binding, and use
them to explore branching code paths that *don't* and point at them. If
we find *no* potential places where an assignment to the binding might
be missing, we display the spans of all the existing initializers to
provide some context.

Of primary interest, this merges
rust-lang/rls@ece09b8 into rust-lang/rust,
which brings in the changes that fix RLS tests broken by rust-lang#97853. rust-lang#97853 already
introduced that commit's changes (under
27f4044df03d15c7c38a483c3e4635cf4f51807d) but without putting those changes on
rust-lang/rls as a branch, so we ended up with an orphan commit that caused
trouble when updating submodules in rust-lang/rust.

This commit, once merged into rust-lang/rust, should continue to let RLS tests
to pass on rust-lang/rust's side and move us back into a healthy state where tip
of the submodule points to a valid master commit in the rust-lang/rls
repository.

RFC-2841: add codegen flag export symbols from executable

Closes rust-lang#84161
r? `@nikomatsakis` `@Mark-Simulacrum`

std: use an event-flag-based thread parker on SOLID

`Mutex` and `Condvar` are being replaced by more efficient implementations, which need thread parking themselves (see rust-lang#93740). Therefore, the generic `Parker` needs to be replaced on all platforms where the new lock implementation will be used, which, after rust-lang#96393, are SOLID, SGX and Hermit (more PRs coming soon).

SOLID, conforming to the [μITRON specification](http://www.ertl.jp/ITRON/SPEC/FILE/mitron-400e.pdf), has event flags, which are a thread parking primitive very similar to `Parker`. However, they do not make any atomic ordering guarantees (even though those can probably be assumed) and necessitate a system call even when the thread token is already available. Hence, this `Parker`, like the Windows parker, uses an extra atomic state variable.

I future-proofed the code by wrapping the event flag in a `WaitFlag` structure, as both SGX and Hermit can share the Parker implementation, they just have slightly different primitives (SGX uses signals and Hermit has a thread blocking API (which is unfortunately [broken](hermit-os/kernel#442), I think).

```@kawadakk``` I assume you are the target maintainer? Could you test this for me?

…ompiler-errors

[rustc_parse] Forbid `let`s in certain places

Currently only forbids in locals to resolve rust-lang#94927 (comment) but feel free to point any other places.

…notriddle

Transform help popup into a pocket menu

Just like we moved the settings menu into a "pocket menu", it's doing the same to the help popup.

You can test it [here](https://rustdoc.crud.net/imperio/help-pocket-menu/doc/foo/index.html) and here is a screenshot:

![Screenshot from 2022-06-20 20-58-29](https://user-images.githubusercontent.com/3050060/174663718-538e9d11-3bf9-48b2-8909-f9bfe75af135.png)

r? ````@jsha````

 On partial uninit error point at where we need init

When a binding is declared without a value, borrowck verifies that all
codepaths have *one* assignment to them to initialize them fully. If
there are any cases where a condition can be met that leaves the binding
uninitialized or we attempt to initialize a field of an uninitialized
binding, we emit E0381.

We now look at all the statements that initialize the binding, and use
them to explore branching code paths that *don't* and point at them. If
we find *no* potential places where an assignment to the binding might
be missing, we display the spans of all the existing initializers to
provide some context.

Fix rust-lang#97956.

…ing, r=oli-obk

Fix printing `impl trait` under binders

Before, we would render `impl for<'a> Trait<'a>` like `impl Trait<for<'a> 'a>`, lol.

…lbini

Bump RLS to latest master on rust-lang/rls

Of primary interest, this merges
rust-lang/rls@ece09b8 into rust-lang/rust,
which brings in the changes that fix RLS tests broken by rust-lang#97853. rust-lang#97853 already
introduced that commit's changes (under
rust-lang/rls@27f4044) but without putting those changes on
rust-lang/rls as a branch, so we ended up with an orphan commit that caused
trouble when updating submodules in rust-lang/rust.

This commit, once merged into rust-lang/rust, should continue to let RLS tests
to pass on rust-lang/rust's side and move us back into a healthy state where tip
of the submodule points to a valid master commit in the rust-lang/rls
repository.

cc rust-lang#98451, but not marking as fixed as I believe we need to add verification to prevent future oversights.

…abort, r=Dylan-DPC

Fix backtrace UI test when panic=abort is used

The function `contains_verbose_expected` is only used when the panic strategy is not abort, so it caused a warning when it was abort, which made the UI test failed on stderr comparison.