proc_macro crate missing from miri sysroot beginning nightly-2023-05-05

[dtolnay]

// src/main.rs

extern crate proc_macro;

fn main() {}

$ cargo +nightly-2023-05-04 miri run  # works

$ cargo +nightly-2023-05-05 miri run
   Compiling testing v0.0.0
    Finished dev [unoptimized + debuginfo] target(s) in 0.11s
     Running `~/.rustup/toolchains/nightly-2023-05-05-x86_64-unknown-linux-gnu/bin/cargo-miri runner target/miri/x86_64-unknown-linux-gnu/debug/testing`

error[E0463]: can't find crate for `proc_macro`
 --> src/main.rs:1:1
  |
1 | extern crate proc_macro;
  | ^^^^^^^^^^^^^^^^^^^^^^^^ can't find crate

error: aborting due to previous error

For more information about this error, try `rustc --explain E0463`.

Rust-lang/rust commit range is rust-lang/rust@473f916...74c4821. Probably rust-lang/rust#108865 (FYI @Zoxc @jyn514)

[jyn514]

Miri probably needs to change from building test to the new sysroot crate.

[saethlin]

Ouch. I might look into this tomorrow, but for anyone else interested in fixing this, Miri uses this crate to build the sysroot: https://github.com/RalfJung/rustc-build-sysroot

[RalfJung]

Yeah... that is the easy part though: RalfJung/rustc-build-sysroot#6. The hard part is that this crate would like to support a range of rustc versions, not just the latest nightly.

[workingjubilee]

I've posted rust-lang/rust#111234 to hopefully give a bit more time to develop a migration plan.

[RalfJung]

For Miri at least I think the migration is done when rust-lang/rust#111236 lands.