fix: Clear polling timer in waitForAvailableWorker on timeout

[bbopen]

Problem

As identified in PR #136 code review, the polling setTimeout in waitForAvailableWorker continues recursively after the timeout rejects, potentially leaking timers.

Current Behavior

private async waitForAvailableWorker(timeoutMs: number = 5000): Promise<WorkerProcess> {
  return new Promise((resolvePromise, reject) => {
    const timeout = setTimeout(() => {
      reject(new Error('Timeout waiting for available worker'));
    }, timeoutMs);

    const checkWorker = (): void => {
      const worker = this.selectOptimalWorker();
      if (worker) {
        clearTimeout(timeout);
        resolvePromise(worker);
      } else {
        setTimeout(checkWorker, 10);  // ← Continues after timeout rejects
      }
    };

    checkWorker();
  });
}

Proposed Fix

Track the polling timer and clear it on timeout:

private async waitForAvailableWorker(timeoutMs: number = 5000): Promise<WorkerProcess> {
  return new Promise((resolvePromise, reject) => {
    let pollTimer: NodeJS.Timeout | null = null;
    let resolved = false;

    const timeout = setTimeout(() => {
      resolved = true;
      if (pollTimer) clearTimeout(pollTimer);
      reject(new Error('Timeout waiting for available worker'));
    }, timeoutMs);

    const checkWorker = (): void => {
      if (resolved) return;
      const worker = this.selectOptimalWorker();
      if (worker) {
        resolved = true;
        clearTimeout(timeout);
        resolvePromise(worker);
      } else {
        pollTimer = setTimeout(checkWorker, 10);
      }
    };

    checkWorker();
  });
}

Impact

Low - The leaked timers resolve to no-ops once the Promise settles (resolved promises ignore subsequent resolve/reject calls). However, this is a code quality issue that should be fixed.

References

• PR refactor: unify NodeBridge and OptimizedNodeBridge (ADR-0001) #136 CodeRabbit review comment
• Related to ADR-0001 bridge unification