This is then later used by `proc_macro` to generate a new
`proc_macro::TokenTree` which preserves span information. Unfortunately this
isn't a bullet-proof approach as it doesn't handle the case when there's still
other attributes on the item, especially inner attributes.

Despite this the intention here is to solve the primary use case for procedural
attributes, attached to functions as outer attributes, likely bare. In this
situation we should be able to now yield a lossless stream of tokens to preserve
span information.

This test currently fails because the tokenization of an AST item during the
expansion of a procedural macro attribute rounds-trips through strings, losing
span information.

The function should accept feature strings that old LLVM might not
support.

Simplify the code using the same approach used by
LLVMRustPrintTargetFeatures.

Dummify the function for non 4.0 LLVM and update the tests accordingly.

One set of tests is to ensure the current message is correct.
The other set is to check the messages are generated correctly for every
type of loop.

Point at the correct path segment on a import statement where a module
doesn't exist.

New output:

```rust
error[E0432]: unresolved import `std::bar`
 --> <anon>:1:10
  |
1 | use std::bar::{foo1, foo2};
  |          ^^^ Could not find `bar` in `std`
```

instead of:

```rust
error[E0432]: unresolved import `std::bar::foo1`
 --> <anon>:1:16
  |
1 | use std::bar::{foo1, foo2};
  |                ^^^^ Could not find `bar` in `std`

error[E0432]: unresolved import `std::bar::foo2`
 --> <anon>:1:22
  |
1 | use std::bar::{foo1, foo2};
  |                      ^^^^ Could not find `bar` in `std`
```

Stabilizes:

* `compile_error!` as a macro defined by rustc

Closes #40872

This commit adds a disabled builder which will run all tests for the standard
library for aarch64 in a QEMU instance. Once we get enough capacity to run this
on Travis this can be used to boost our platform coverage of AArch64

Fixes issue #43317.

This is in the matter of #35164.

Float literals need to be parsed as the correct type so they can be
rounded correctly.