Some platforms don't actually have `libm` already linked in the test
infrastructure, and then `dynamic_lib::tests::test_loading_cosine` would
fail to find the "cos" symbol.  Every platform running this test should
have `libc` and "atoi" though, so try to use that symbol instead.

Fixes rust-lang#45410.

Previously, using unknown as the vendor value would lead to the same
result, but with the multiarch runtimes support in Clang, the target is
now used to locate the runtime libraries and so the format is important.
The denormalized format with omitted vendor component is the format we
use with Clang and should be using for Rust as well.

…matsakis

Do a basic sanity check for all constant values

## Motivation and high level overview

There has been some back and forth in this PR between @RalfJung and me in here about the motivation for this change and the stance it takes on unsafe coding guidelines.

The initial implementation ran its checks on every value read (so `*x`, `y = x`, ...). In unsafe code that isn't reasonable, because we might be invalidating invariants for a short time in order to build up a proper value.

The current implementation is a lint that runs its checks statics and constants. There is no need to check array lengths and enum variants, because it's a hard error to end up with anything but a number, and that one even has to have the required bits to be defined.

## What checks are done?

* Some type related checks
    * `char` needs to be a correct unicode character as defined by `char::from_u32`
    * A reference to a ZST must have the correct alignment (and be nonzero)
    * A reference to anything is dereferenced and its value is checked
* Layout checks use the information from `ty::Layout` to check
    * all fields of structs
    * all elements of arrays
    * enum discriminants
    * the fields of an enum variant (the variant is decided by the discriminant)
    * whether any union field succeeds in being checked (if none match the memory pattern, the check fails)
    * if the value is in the range described by the layout (e.g. for `NonZero*` types)

Changing the layout of a type will thus automatically cause the checks to check for the new layout.

fixes rust-lang#51330
fixes rust-lang#51471

cc @RalfJung

r? @eddyb

Suggest fix when encountering different mutability from impl to trait

Closes rust-lang#52412
r? @estebank

…hton

Minor improvements to bootstrap

- prefer `Path`-specific methods to `String` ones
- don't add file extensions if they are removed right afterwards

make memrchr use align_offset

I hope I did not screw that up...

Cc @oli-obk who authored the original rust-lang#44537

… r=GuillaumeGomez

rustdoc: rework how default passes are chosen

This is a refactor that changes how we select default passes, and changes the set of passes used for `--document-private-items`. It's groundwork for a bigger refactor i want to do.

The major changes:

* There are now two sets of "default passes": one set for "no flags given" and one for "document private items".
* These sets can be selected by a new `DefaultPassOption` enum, which is selected from based on the presence of `--no-defaults` or `--document-private-items` CLI flags, or their associated crate attributes.
* When printing the list of passes, we also print the list of passes for `--document-private-items` in addition to the "default defaults".
* I added `propagate-doc-cfg` and `strip-priv-imports` to the "document private items" set. The former is to ensure items are properly tagged with the full set of cfg flags even when "document private items" is active. The latter is based on feedback and personal experience navigating the `rustc` docs, which use that flag. `strip-priv-imports` only removes non-pub `use` statements, so it should be harmless from a documentation standpoint to remove those items from "private items" documentation.

…dle, r=TimNN

Impl Send & Sync for JoinHandle

This is just a cosmetic change - it slightly relaxes and clarifies the public API without effectively promising any new guarantees.

Currently we have [these auto trait implementations](https://doc.rust-lang.org/nightly/std/thread/struct.JoinHandle.html#synthetic-implementations):

```rust
impl<T: Send> Send for JoinHandle<T> {}
impl<T: Sync> Sync for JoinHandle<T> {}
```

Bound `T: Send` doesn't make much sense because `JoinHandle<T>` can be created only when `T: Send`. Note that [`JoinHandle::<T>::join`](https://doc.rust-lang.org/nightly/std/thread/struct.JoinHandle.html#method.join) doesn't require `T: Send` so why should the `Send` impl?

And the `Sync` impl doesn't need `T: Sync` because `JoinHandle<T>` cannot even share `T` - it can only send it to the thread that calls `join`.

…ichton

rustc_metadata: test loading atoi instead of cos

Some platforms don't actually have `libm` already linked in the test
infrastructure, and then `dynamic_lib::tests::test_loading_cosine` would
fail to find the "cos" symbol.  Every platform running this test should
have `libc` and "atoi" though, so try to use that symbol instead.

Fixes rust-lang#45410.

…chton

Omit the vendor component in Fuchsia triple

Previously, using unknown as the vendor value would lead to the same
result, but with the multiarch runtimes support in Clang, the target is
now used to locate the runtime libraries and so the format is important.
The denormalized format with omitted vendor component is the format we
use with Clang and should be using for Rust as well.

…r=pnkfelix

Remove unused "-Zenable_nonzeroing_move_hints" flag

Removing a dead option which seems to be a remnant of the old drop-flag system.

Incorporate a stray test

`liballoc/repeat-generic-slice.rs` doesn't seem to be tested (I think it was intended to be placed in `run-pass`). This PR incorporates the test into `liballoc/tests`.

Fix doc comment for 'ptr::copy_to' method

Fix error in doc comment for `ptr::copy_to` method.

revert accidental atty downgrade

This got accidentally downgraded by rust-lang#52488

Cc @nikomatsakis @pnkfelix