Function worker_index 

pub fn worker_index() -> usize

Returns the zero-based worker index assigned to this OS process.

This is a process-level identifier, not a per-test-thread identifier. Each spawned worker subprocess has at most one index; the top-level parent and any “no spawn workers” execution path (e.g. --nocapture, no captured-output requirement, or Shared deps forcing single-thread execution) observes 0.

When this is useful

Only PerWorker constructors and the tests they feed see meaningful per-worker values, because they are the only constructors that run inside spawned worker subprocesses. Parent-only scopes — Shared, Cloneable, Hosted, and HostedRpc — always run in the top-level parent and therefore always observe 0. Do not use worker_index() to partition state inside those constructors; the answer is always 0.

Use with PerWorker dependencies

Combine with #[test_dep(scope = PerWorker)] to seed per-worker state. For example, a unique-id counter that partitions its id space by worker so that two parallel workers cannot mint the same id:

use std::sync::atomic::AtomicU16;
use test_r::test_dep;

pub struct LastUniqueId {
    pub id: AtomicU16,
}

#[test_dep(scope = PerWorker)]
fn last_unique_id() -> LastUniqueId {
    // Reserve the high 8 bits for the worker index, leaving 8 bits per
    // worker for the local sequence. Adjust the shift to whatever fits
    // the underlying integer width.
    LastUniqueId {
        id: AtomicU16::new((test_r::worker_index() as u16) << 8),
    }
}