WorkerPool allows zero workers but next_worker() panics on an empty pool

[xc01]

Describe the bug

WorkerPool::new(..., 0) is currently allowed and returns Ok(...):

/// This can happen if the pool was created with 0 workers
/// Which is not particularly useful, but is allowed

However, calling next_worker() on that pool panics because it indexes workers[0] even when the pool is empty:

let worker = &self.workers[self.next_worker];

This means a configuration value of 0 workers is accepted during initialization, but later turns into a runtime panic when work is dispatched.

In a server setting, this can move a configuration error from startup time into request/task handling.

To Reproduce (PoC)

use rustyscript::worker::{DefaultWorker, DefaultWorkerOptions, WorkerPool};

#[test]
#[should_panic]
fn worker_pool_next_worker_panics_when_pool_is_empty() {
    let mut pool = WorkerPool::<DefaultWorker>::new(
        DefaultWorkerOptions::default(),
        0,
    )
    .expect("creating a zero-worker pool is currently allowed");

    let _ = pool.next_worker();
}

Expected behavior

Either:

• WorkerPool::new(..., 0) should return an error, or
• next_worker() should gracefully handle an empty pool instead of panicking.

Desktop

OS: Ubuntu 22.04.1
Version: 0.12.3