Testing & Code Quality

Testing Async Code

Async tests must wait for the work they assert on. Promise returns, async functions, fake timers, microtasks, cancellation, and unhandled rejections all affect test reliability.

Return or await the async work
Rejection paths need explicit assertions
Fake timers control scheduled work
Microtasks and timers are different queues
Unhandled rejections should fail tests

Async success and failure

it("loads user", async () => {
  await expect(loadUser("u1")).resolves.toMatchObject({ id: "u1" })
})

it("rejects invalid id", async () => {
  await expect(loadUser("")).rejects.toThrow("Invalid")
})

The test awaits both success and failure assertions.

Fake timers for debounce

import { afterEach, expect, it, vi } from "vitest"

afterEach(() => vi.useRealTimers())

it("debounces search", () => {
  vi.useFakeTimers()
  const fn = vi.fn()
  const search = debounce(fn, 300)

  search("a")
  search("ab")
  vi.advanceTimersByTime(299)
  expect(fn).not.toHaveBeenCalled()

  vi.advanceTimersByTime(1)
  expect(fn).toHaveBeenCalledWith("ab")
})

The test controls time instead of waiting 300 real milliseconds.