This commit makes the `clean::Type::is_same` non-commutative, so
that a generic `impl` matches a concrete return, but a generic return
does not match a concrete `impl`. It makes slice and vector Write
for `u8` not match on every generic return value.

Bring the the `#![allow(internal_features)]`

…generic, r=camelid

rustdoc: handle generics better when matching notable traits

This commit makes the `clean::Type::is_same` non-commutative (renaming it `is_doc_subtype_of`), so that a generic `impl` matches a concrete return, but a generic return does not match a concrete `impl`. It makes slice and vector Write for `u8` not match on every generic return value.

Fixes rust-lang#100322

Fixes rust-lang#55082

Preview:

* https://notriddle.com/rustdoc-demo-html-3/notable-trait-generic/std/vec/struct.Vec.html#method.new
* https://notriddle.com/rustdoc-demo-html-3/notable-trait-generic/std/vec/struct.Vec.html#method.from-12
* https://notriddle.com/rustdoc-demo-html-3/notable-trait-generic/std/iter/trait.Iterator.html#method.intersperse_with
* https://notriddle.com/rustdoc-demo-html-3/notable-trait-generic/std/string/struct.String.html#method.as_bytes

…=Nilstrieb

refactor/feat: refactor identifier parsing a bit

\+ error recovery for `expected_ident_found`

Prior art: rust-lang#108854

Eagerly intern and check CrateNum/StableCrateId collisions

r? ``@cjgillot``

It seems better to check things ahead of time than checking them afterwards.
The [previous version](rust-lang#108390) was a bit nonsensical, so this addresses the feedback

rustc: Remove unused `Session` argument from some attribute functions

(One auxiliary test file containing one of these functions was unused, so I removed it instead of updating.)

Update stdarch

Bring the the `#![allow(internal_features)]` for rust-lang#108955

r? `@Amanieu`

Remove Ty::is_region_ptr

Fixes rust-lang#109372

…-in-self-ty, r=compiler-errors

Use region-erased self type during IAT selection

Split off from rust-lang#109410 as discussed.
Fixes rust-lang#109299.

Re UI test: I use a reproducer of rust-lang#109299 that contains a name resolution error instead of reproducer [`regionck-2.rs`](https://github.com/rust-lang/rust/blob/fc7ed4af165c27ab5914b93251194f826920cc65/tests/ui/associated-inherent-types/regionck-2.rs) (as found in the `AliasKind::Inherent` PR) since it would (incorrectly) pass typeck in this PR due to the lack of regionck and I'd rather not make *that* a regression test (with or without `known-bug`).

``@rustbot`` label F-inherent_associated_types

r? ``@compiler-errors``

new solver cleanup + implement coherence

the cleanup:
- change `Certainty::unify_and` to consider ambig + overflow to be ambig
- rename `trait_candidate_should_be_dropped_in_favor_of` to `candidate_should_be_dropped_in_favor_of`
- remove outdated fixme

For coherence I mostly just add an ambiguous candidate if the current trait ref is unknowable. I am doing the same for reservation impl where I also just add an ambiguous candidate.

make link clickable