Make the following API stable:

    // in core::any
    pub fn type_name_of_val<T: ?Sized>(_val: &T) -> &'static str

Const stability is not added because this relies on `type_name` which is also
not const. That has a blocking issue.

Fixes rust-lang#66359

I've been inactive around diagnostics for a while so would like to remove myself from the diagnostics-related ping groups for now.

Follow up for rust-lang#118961

…dtolnay

Stabilize `ptr::{from_ref, from_mut}`

I propose to stabilize the following APIs:

```rust
// mod core::ptr

pub const fn from_ref<T: ?Sized>(r: &T) -> *const T;
pub const fn from_mut<T: ?Sized>(r: &mut T) -> *mut T;
```

Tracking issue: rust-lang#106116

---

`@RalfJung` what do you think we should do with `from_mut`? Stabilize it as const, given that you can't call it anyway (no way to get `&mut` in `const fn`)? Defer stabilizing it as const leaving the same issue/feature? Change issue/feature? Change issue/feature to the "`&mut` in const fn" one?

…lnay

Stabilize `type_name_of_val`

Make the following API stable:

```rust
// in core::any
pub fn type_name_of_val<T: ?Sized>(_val: &T) -> &'static str
```

This is a convenience method to get the type name of a value, as opposed to `type_name` that takes a type as a generic.

Const stability is not added because this relies on `type_name` which is also not const. That has a blocking issue rust-lang#97156.

Wording was also changed to direct most of the details to `type_name` so we don't have as much duplicated documentation.

Fixes tracking issue rust-lang#66359.

There were two main concerns in the tracking issue:

1. Naming: `type_name_of` and `type_name_of_val` seem like the only mentioned options. Differences in opinion here come from `std::mem::{size_of, align_of, size_of_val, align_of_val}`. This PR leaves the name as `type_name_of_val`, but I can change if desired since it is pretty verbose.
2. What this displays for `&dyn`: I don't think that having `type_name_of_val` function resolve those is worth the headache it would be, see rust-lang#66359 (comment) for some workarounds. I also amended the docs wording to leave it open-ended, in case we have means to change that behavior in the future.

`@rustbot` label -T-libs +T-libs-api +needs-fcp
r? libs-api

Move type relations into submodule `relate` in rustc_infer, and notify when it has changed

r? lcnr

… r=notriddle

Simplify `src-script.js` code

Instead of keeping this value in the global scope and still use it in the function in case it wasn't used outside, let's just use it inside the function.

r? `@notriddle`

…compiler-errors

Remove @JohnTitor from diagnostics pings

I've been inactive around diagnostics for a while so would like to remove myself from the diagnostics-related ping groups for now.

… r=notriddle

Simplify JS code a little bit

As mentioned, JS code can be simplified a little bit.

r? `@notriddle`

…illaumeGomez

rustdoc: add regression test for JS data file loading

Follow up for rust-lang#118961