/**
 * Per-scanner run boundary for tests that call a scanner's `scan()` directly.
 *
 * `scan-orchestrator` establishes this boundary before every scanner; a test that
 * calls `scan()` twice in one process otherwise carries state across the two runs
 * and confounds what it measures (M-BUG-28, session #59).
 *
 * Finding IDs are now derived from the check registry rather than from emission
 * order, so there is no per-run state left and this IS a deliberate no-op. It is
 * kept as the seam: the tests that call it assert the same thing either way, and
 * that is the point — the assertion never depended on the counter.
 */

export function startScannerRun() {
  // Intentionally empty: `finding()` holds no per-run state.
}