break rustc_expand-rustc_middle dependency

[mejrs]

Back in #145354 (cc @Kobzol), support for caching derive macros was added. With this, rustc_expand was made to depend on rustc_middle.

This PR breaks that dependency so that rustc_expand and rustc_builtin_macros can compile in parallel with rustc_middle.

The timing graph goes from this:

to this:

Note that the interval where we're exclusively compiling rustc_middle has become much shorter.
Full graphs here:
cargo-timing-main.html
cargo-timing-expand2.html

It's somewhat hard to benchmark reliably but a full bootstrap is about 5-10 seconds faster overall, on a machine with an Amd Ryzen 5900x 12-core processor.

[rustbot]

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

Some changes occurred in src/tools/cargo

cc @weihanglo

[rustbot]

r? @JonathanBrouwer

rustbot has assigned @JonathanBrouwer.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 75 candidates
• Random selection from 15 candidates

[mejrs]

Shouldn't show up in single job benchmarks, but still:

I'll also do another benchmark later where the derive expander is always called through the atomic ref (with enable called depending on whether derive macros are cached or not)

@bors try @rust-timer queue

[rust-timer]

Awaiting bors try build completion.

@rustbot label: +S-waiting-on-perf

[rust-bors]

☀️ Try build successful (CI)
Build commit: e3517d8 (e3517d88b373f392501ed15f89f9f57cd5370ddf)
Base parent: e702eca (e702ecae8a3e3756e8a499f8592be1e3efac0e8f)

[rust-timer]

Queued e3517d8 with parent e702eca, future comparison URL.
There is currently 1 preceding artifact in the queue.
It will probably take at least ~1.7 hours until the benchmark run finishes.