Challenge 2: Testing an Inactivity Timer — Possible Solution
====================================================================

it("shows the inactivity prompt after 60 seconds", () => {
  jest.useFakeTimers();
  render(<InactivityPrompt />);

  expect(screen.queryByText("Are you still there?")).not.toBeInTheDocument();

  act(() => {
    jest.advanceTimersByTime(60000);
  });

  expect(screen.getByText("Are you still there?")).toBeInTheDocument();
});

WHY THIS WORKS AS AN ANSWER
------------------------------
jest.useFakeTimers() replaces the real setTimeout/setInterval
implementations with controllable fake ones BEFORE the component
renders, so the component's internal timer (whatever triggers the
60-second inactivity check) never actually waits in real time at all.

The first assertion uses queryByText (not getByText) because the
message is genuinely expected to be ABSENT immediately after render —
zero real or fake time has passed yet — exactly Chapter 3's rule about
when queryBy is required instead of getBy.

jest.advanceTimersByTime(60000) instantly advances the fake clock by
60,000 milliseconds (60 seconds) — wrapped in act(...) because this
directly triggers a state update inside the component (the timer
firing), matching this chapter's own toast-dismissal example's use of
act() around the same function. Without wrapping it in act(), the
following assertion could read a stale render, per Chapter 5's act()
explanation.

After advancing time, getByText (not queryByText) is now appropriate,
since the message SHOULD exist at this point — using getBy here
correctly asserts a positive expectation rather than checking for
absence.