For a given file

```rust
trait A { fn foo(&self) {} }
trait B : A { fn foo(&self) {} }

fn bar<T: B>(a: &T) {
  a.foo()
}
```

provide the following output

```
error[E0034]: multiple applicable items in scope
 --> file.rs:6:5
  |
6 |   a.foo(1)
  |     ^^^ multiple `foo` found
  |
note: candidate #1 is defined in the trait `A`
 --> file.rs:2:11
  |
2 | trait A { fn foo(&self, a: usize) {} }
  |           ^^^^^^^^^^^^^^^^^^^^^^^^^^
help: to use it here write `A::foo(&a, 1)` instead
 --> file.rs:6:5
  |
6 |   a.foo(1)
  |     ^^^
note: candidate #2 is defined in the trait `B`
 --> file.rs:3:15
  |
3 | trait B : A { fn foo(&self, a: usize) {} }
  |               ^^^^^^^^^^^^^^^^^^^^^^^^^^
help: to use it here write `B::foo(&a, 1)` instead
 --> file.rs:6:5
  |
6 |   a.foo(1)
  |     ^^^
```

For legacy reasons (presumably), Windows does not permit files name aux.

Instead of finding aux-build files in `auxiliary`, we now search for an
`aux` directory relative to the test. So if your test is
`compile-fail/foo.rs`, we would look in `compile-fail/aux`.  Similarly,
we ignore the `aux` directory when searching for tets.

LLVM's memory dependence analysis doesn't properly account for calls
that could unwind and thus effectively act as a branching point. This
can lead to stores that are only visible when the call unwinds being
removed, possibly leading to calls to drop() functions with b0rked
memory contents.

As there is no fix for this in LLVM yet and we want to keep
compatibility to current LLVM versions anyways, we have to workaround
this bug by omitting the noalias attribute on &mut function arguments.
Benchmarks suggest that the performance loss by this change is very
small.

Thanks to @RalfJung for pushing me towards not removing too many
noalias annotations and @alexcrichton for helping out with the test for
this bug.

Fixes #29485

Fixes #29857.
Fixes #30589.

macro future proofing rules.

(We may want to think about what this test was actually testing and
figure out a way to test it without running afoul of macro future
proofing.  I spent some time trying to do this, e.g. by inserting
parenthesis in the macro input pattern, but I could not quickly get it
working, so I took this tack instead.)

Fixes #30715

Now that features must be declared expanded source often does not compile.
This adds 'pretty-expanded' to a bunch of test cases that still work.

This breaks code that referred to variant names in the same namespace as
their enum. Reexport the variants in the old location or alter code to
refer to the new locations:

```
pub enum Foo {
    A,
    B
}

fn main() {
    let a = A;
}
```
=>
```
pub use self::Foo::{A, B};

pub enum Foo {
    A,
    B
}

fn main() {
    let a = A;
}
```
or
```
pub enum Foo {
    A,
    B
}

fn main() {
    let a = Foo::A;
}
```

[breaking-change]

RFC: 0008-match-arm-attributes

This is required by the check-fast target because each test is slurped up into a
submodule.