[CLAUDE ROUTINE]: Test-coverage enhancement — wrap vi.useFakeTimers/useRealTimers in try/finally (or afterEach) to prevent leaking fake timers across tests #173

Summary

__tests__/components/copy-button.test.tsx enables fake timers inside a test body and restores real timers at the end of the same body. If the test throws between the two calls (an assertion failure, an unhandled promise rejection, …), real timers are never restored and every subsequent test in the same worker runs with fake timers still active — the classic "one red test turns the rest red" flakiness pattern.

Where

__tests__/components/copy-button.test.tsx (around the it(... 'shows copied state for 1.5s' ...) block)

vi.useFakeTimers({ shouldAdvanceTime: true });
render(<CopyButton text="test" />);
// ← if this block throws, real timers are never restored
vi.useRealTimers();

Why this helps users

• Clean CI signal — a single flake stops cascading into spurious failures in later tests, making the test job (one of the four CI jobs) more trustworthy.
• Faster triage — when a timer assertion fails, the next test's unrelated error no longer distracts the person on-call.
• Scales to [CLAUDE ROUTINE]: Test-coverage enhancement — add unit tests for src/auth/ and src/relay/ modules #155 / [CLAUDE ROUTINE]: Test-coverage enhancement — per-policy tests for built-in policies with edge cases and params #156 / [CLAUDE ROUTINE]: Test-coverage enhancement — loader edge cases (symlinks, deep transitive imports, timeout cleanup) #157 — those issues will add more timer-sensitive suites; establishing the pattern now pays forward.

Proposed enhancement

Prefer a global hook so individual tests don't need to remember:

// __tests__/setup.ts (or a local beforeEach/afterEach in the suite)
afterEach(() => {
  vi.useRealTimers();
  vi.clearAllTimers();
});

For the in-test usage, wrap in try/finally:

vi.useFakeTimers({ shouldAdvanceTime: true });
try {
  render(<CopyButton text="test" />);
  // …
} finally {
  vi.useRealTimers();
}

Acceptance criteria

• vi.useFakeTimers callsites are either wrapped in try/finally or relying on an afterEach reset.
• Deliberately-failing timer test does not cause unrelated follow-on tests to go red.
• CHANGELOG.md entry under Unreleased → Docs (test-only change).

Severity

Low — no production impact; test-suite hygiene tune-up.