Custom MIR: Support binary and unary operations
Lower binary and unary operations directly to corresponding unchecked MIR
operations. Ultimately this might not be syntax we want, but it allows for
experimentation in the meantime.
r? ````@oli-obk```` ````@JakobDegen````
- Eliminates all the `get_context` calls that async lowering created.
- Replace all `Local` `ResumeTy` types with `&mut Context<'_>`.
The `Local`s that have their types replaced are:
- The `resume` argument itself.
- The argument to `get_context`.
- The yielded value of a `yield`.
The `ResumeTy` hides a `&mut Context<'_>` behind an unsafe raw pointer, and the
`get_context` function is being used to convert that back to a `&mut Context<'_>`.
Ideally the async lowering would not use the `ResumeTy`/`get_context` indirection,
but rather directly use `&mut Context<'_>`, however that would currently
lead to higher-kinded lifetime errors.
See <https://github.com/rust-lang/rust/issues/105501>.
The async lowering step and the type / lifetime inference / checking are
still using the `ResumeTy` indirection for the time being, and that indirection
is removed here. After this transform, the generator body only knows about `&mut Context<'_>`.
The test cases src/test/mir-opt/building/custom/consts.rs and
src/test/mir-opt/const_prop/mutable_variable_no_prop.rs are
currently failing on big-endian platforms as the binary encoding
of some constants is hard-coded in the MIR test files. Fix this
by choosing constant values that have the same encoding on big-
and little-endian platforms.
The test case src/test/mir-opt/issues/issue_75439.rs is failing
as well, but since the purpose of the test is to validate handling
of big-endian integer encodings on a little-endian platform, it does
not make much sense to run it on big-endian platforms in the first
place - we can just ignore it there.
Fixed part of https://github.com/rust-lang/rust/issues/105383.
