The existing wording was inappropriate for e.g.
`if let Ok(_) = expr { .. }`. The diagnostic would
leak the fact that we desugar to a `match`.

TODO: The type is wrong and will be fixed in later commits.

Current name is too specific for incoming changes.

when type checking `<pat>`.

…lim"

LLVM 8 (D56351) introduced "frame-pointer". In LLVM 10 (D71863),
"no-frame-pointer-elim"/"no-frame-pointer-elim-non-leaf" will be
ignored.

…imulacrum

Extract `rustc_ast_lowering` crate from `rustc`

Working towards rust-lang#65031.

This PR moves `src/librustc/hir/lowering{/, .rs}` to its own crate (`librustc_ast_lowering`) which is very self-contained (only `fn lower_crate` and `trait Resolver` are exposed).

r? @Mark-Simulacrum

…henkov

Add symbol normalization for proc_macro_server.

Follow up for rust-lang#66670, finishing the first bullet point in rust-lang#55467.

r? @petrochenkov

Cleanup pattern type checking, fix diagnostics bugs (+ improvements)

r? @estebank

…enkov

parser: reduce diversity in error handling mechanisms

Instead of having e.g. `span_err`, `fatal`, etc., we prefer to move towards uniformly using `struct_span_err` thus making it harder to emit fatal and/or unstructured diagnostics.

This PR also de-fatalizes some diagnostics.

r? @estebank

Use function attribute "frame-pointer" instead of "no-frame-pointer-elim"

LLVM 8 ([D56351](http://reviews.llvm.org/D56351)) introduced "frame-pointer". In LLVM 10 (D71863),
"no-frame-pointer-elim"/"no-frame-pointer-elim-non-leaf" will be
ignored.

-----

In the LLVM monorepo, run `git show origin/release/8.x:llvm/lib/CodeGen/TargetOptionsImpl.cpp` to see that `"frame-pointer"` is available since LLVM 8.