Fixes rust-lang#61863

We now allow uses of 'existential type's that aren't defining uses -
that is, uses which don't constrain the underlying concrete type.

To make this work correctly, we also modify eq_opaque_type_and_type to
not try to apply additional constraints to an opaque type. If we have
code like this:

```
existential type Foo;
fn foo1() -> Foo { ... }
fn foo2() -> Foo { foo1() }
```

then 'foo2' doesn't end up constraining 'Foo', which means that
'foo2' will end up using the type 'Foo' internally - that is, an actual
'TyKind::Opaque'. We don't want to equate this to the underlying
concrete type - we just need to enforce the basic equality constraint
between the two types (here, the return type of 'foo1' and the return
type of 'foo2')

Previously, types like (Foo, u8) would not be handled correctly
(where Foo is an 'existential type')

Co-Authored-By: Mazdak Farrokhzad <twingoow@gmail.com>

This commit adds a reproduction of the error reported in servo which
demonstrates the current, incorrect behaviour.

Co-authored-by: Rémy Rakić <remy.rakic@gmail.com>

The Dumper no longer has anything to do specifically with JSON, it
merely represents processing into an `Analysis` output.

Co-Authored-By: Ralf Jung <post@ralfj.de>

Co-Authored-By: Ralf Jung <post@ralfj.de>

This commit implements a hacky fix for detecting when a span is pointing
at a macro call site so that bare trait lints are not made incorrectly.

…oli-obk

Fix cycle error with existential types

Fixes rust-lang#61863

We now allow uses of `existential type`'s that aren't defining uses - that is, uses which don't constrain the underlying concrete type.

To make this work correctly, we also modify `eq_opaque_type_and_type` to not try to apply additional constraints to an opaque type. If we have code like this:

```rust
existential type Foo;
fn foo1() -> Foo { ... }
fn foo2() -> Foo { foo1() }
```

then `foo2` doesn't end up constraining `Foo`, which means that `foo2` will end up using the type `Foo` internally - that is, an actual `TyKind::Opaque`. We don't want to equate this to the underlying concrete type - we just need to enforce the basic equality constraint between the two types (here, the return type of `foo1` and the return type of `foo2`)

… r=Xanewok

Cleanup save-analysis JsonDumper

Don't access a static just for its size and alignment

cc @RalfJung

fixes rust-lang#62189

add `repr(transparent)` to `IoSliceMut` where missing

tried using `IoSliceMut` in FFI, got `improper_ctypes` warning.

according to the docs: `IoSliceMut` is  "guaranteed to be ABI compatible with the `iovec` type" so it should be usable in FFI.
`IoSlice` is also `repr(transparent)` for every platform where these types contain `iovec`-like types.
vxworks also has `IoSliceMut` as transparent so its not even consistently one or the other.

no comment about this next to the types or in the PR that introduced the types, so assuming this was just missed.

r? @sfackler

…gestion, r=estebank

Stop bare trait lint applying to macro call sites

Fixes rust-lang#61963. Apologies for the delay with in fixing this. If anyone has a better idea how to detect this macro call site case, I'd be happy to fix this in a more robust, less hacky way.

r? @estebank

Add lib section to rustc_lexer's Cargo.toml

This is required to fix the rustc-ap-syntax build error in the recent version. The error could also be fixed on the [rustc-auto-publish](https://github.com/alexcrichton/rustc-auto-publish) side by manually adding `[lib]` section if one does not exist. The latter approach, however, may have a surprising side effect, so I am opting for a simpler solution for now.

r? @alexcrichton