front-end-testing（前端测试）

Front-End Testing

For React-specific patterns (components, hooks, context), load the react-testing skill. For TDD workflow, load the tdd skill. For general testing patterns (factories, public API testing), load the testing skill.

Vitest Browser Mode (Preferred)

Always prefer Vitest Browser Mode over jsdom/happy-dom. Tests run in a real browser (via Playwright), giving production-accurate behavior for CSS, events, focus management, and accessibility.

Why Browser Mode Over jsdom

Setup

npm install -D vitest @vitest/browser-playwright

// vitest.config.ts
import { defineConfig } from 'vitest/config'
import { playwright } from '@vitest/browser-playwright'

export default defineConfig({
  test: {
    browser: {
      enabled: true,
      provider: playwright(),
      headless: true,
      instances: [{ browser: 'chromium' }],
    },
  },
})

Quick setup wizard: npx vitest init browser

Built-in Locators

Vitest Browser Mode has built-in locators that mirror Testing Library queries. No separate @testing-library/dom import needed.

import { page } from 'vitest/browser'

// These work exactly like Testing Library queries
page.getByRole('button', { name: /submit/i })
page.getByText(/welcome/i)
page.getByLabelText(/email/i)
page.getByPlaceholder(/search/i)
page.getByAltText(/logo/i)
page.getByTestId('my-element')  // Last resort only

Built-in Assertions with Retry

Use expect.element() for DOM assertions — it automatically retries until the assertion passes or times out, reducing flakiness:

// ✅ CORRECT - Auto-retrying assertion
await expect.element(page.getByText(/success/i)).toBeVisible()
await expect.element(page.getByRole('button')).toBeDisabled()

// Available matchers (no @testing-library/jest-dom needed):
await expect.element(el).toBeVisible()
await expect.element(el).toBeDisabled()
await expect.element(el).toHaveTextContent(/text/i)
await expect.element(el).toHaveValue('input value')
await expect.element(el).toHaveAttribute('aria-label', 'Close')
await expect.element(el).toBeChecked()

Built-in User Events (CDP-based)

import { userEvent } from 'vitest/browser'

// Real browser events via Chrome DevTools Protocol
await userEvent.click(page.getByRole('button', { name: /submit/i }))
await userEvent.fill(page.getByLabelText(/email/i), 'test@example.com')
await userEvent.keyboard('{Enter}')
await userEvent.selectOptions(page.getByLabelText(/country/i), 'USA')
await userEvent.clear(page.getByLabelText(/search/i))

Or use locator methods directly:

await page.getByRole('button', { name: /submit/i }).click()
await page.getByLabelText(/email/i).fill('test@example.com')

Multi-Project Setup (Node + Browser)

When you need both unit tests (Node) and UI tests (browser):

export default defineConfig({
  test: {
    projects: [
      {
        test: {
          include: ['tests/unit/**/*.test.ts'],
          name: 'unit',
          environment: 'node',
        },
      },
      {
        test: {
          include: ['tests/browser/**/*.test.ts'],
          name: 'browser',
          browser: {
            enabled: true,
            provider: playwright(),
            instances: [{ browser: 'chromium' }],
          },
        },
      },
    ],
  },
})

Browser Mode Gotchas

• vi.spyOn on imports: ES module namespaces are sealed in real browsers. Use vi.mock('./module', {spy: true}) instead.
• alert()/confirm(): Thread-blocking dialogs halt browser execution. Mock them with vi.spyOn(window, 'alert').mockImplementation(() => {}).
• act() not needed: CDP events + expect.element() retry handle timing automatically.

Playwright / Browser Mode Test Idempotency

All Playwright-style tests MUST be idempotent. Every test must produce the same result regardless of execution order, how many times it runs, or what other tests ran before it.

Rules:

• Each test creates its own state from scratch — never depend on another test's side effects
• Clean up any persistent state (database rows, localStorage, cookies) created during the test
• Use unique identifiers (e.g., timestamp-based) to avoid collisions when tests run in parallel
• Never assume the DOM is in a particular state at the start of a test — render fresh
• If tests share a server or database, use isolation strategies (transactions, test-specific data)

// ❌ WRONG - Tests depend on shared state
it('creates a user', async () => {
  await page.getByRole('button', { name: /create/i }).click()
  // Creates user "Alice" in the database
})

it('lists users', async () => {
  // Assumes "Alice" exists from previous test!
  await expect.element(page.getByText('Alice')).toBeVisible()
})

// ✅ CORRECT - Each test is self-contained
it('creates and displays a user', async () => {
  const uniqueName = `User-${Date.now()}`
  await page.getByLabelText(/name/i).fill(uniqueName)
  await page.getByRole('button', { name: /create/i }).click()
  await expect.element(page.getByText(uniqueName)).toBeVisible()
})

Why this matters: Browser Mode can run tests in parallel across multiple browser instances. Non-idempotent tests will produce flaky failures that are nearly impossible to debug.

Legacy: DOM Testing Library Patterns

The patterns below apply when using @testing-library/dom directly (e.g., with jsdom). Prefer Vitest Browser Mode for new projects — the query patterns are identical but built-in.

Core Philosophy

Test behavior users see, not implementation details.

Testing Library exists to solve a fundamental problem: tests that break when you refactor (false negatives) and tests that pass when bugs exist (false positives).

Two Types of Users

Your UI components have two users:

1. End-users: Interact through the DOM (clicks, typing, reading text)
2. Developers: You, refactoring implementation

Kent C. Dodds principle: "The more your tests resemble the way your software is used, the more confidence they can give you."

Why This Matters

False negatives (tests break on refactor):

// ❌ WRONG - Testing implementation (will break on refactor)
it('should update internal state', () => {
  const component = new CounterComponent();
  component.setState({ count: 5 }); // Coupled to state implementation
  expect(component.state.count).toBe(5);
});

False positives (bugs pass tests):

// ❌ WRONG - Testing wrong thing
it('should render button', () => {
  render('<button data-testid="submit-btn">Submit</button>');
  expect(screen.getByTestId('submit-btn')).toBeInTheDocument();
  // Button exists but onClick is broken - test passes!
});

Correct approach (behavior-driven):

// ✅ CORRECT - Testing user-visible behavior
it('should submit form when user clicks submit', async () => {
  const handleSubmit = vi.fn();
  const user = userEvent.setup();

  render(`
    <form id="login-form">
      <label>Email: <input name="email" /></label>
      <label>Password: <input name="password" type="password" /></label>
      <button type="submit">Submit</button>
    </form>
  `);

  document.getElementById('login-form').addEventListener('submit', (e) => {
    e.preventDefault();
    handleSubmit(new FormData(e.target));
  });

  await user.type(screen.getByLabelText(/email/i), 'test@example.com');
  await user.type(screen.getByLabelText(/password/i), 'password123');
  await user.click(screen.getByRole('button', { name: /submit/i }));

  expect(handleSubmit).toHaveBeenCalled();
});

This test:

• Survives refactoring (state → signals → stores)
• Tests the contract (what users see)
• Catches real bugs (broken onClick, validation errors)

Query Selection Priority

Most critical Testing Library skill: choosing the right query.

Priority Order

Use queries in this order (accessibility-first):

1. getByRole - Highest priority

- Queries by ARIA role + accessible name - Mirrors screen reader experience - Forces semantic HTML

1. getByLabelText - Form fields

- Finds inputs by associated <label> - Ensures accessible forms

1. getByPlaceholderText - Fallback for inputs

- Only when label not present - Placeholder shouldn't replace label

1. getByText - Non-interactive content

- Headings, paragraphs, list items - Content users read

1. getByDisplayValue - Current form values

- Inputs with pre-filled values

1. getByAltText - Images

- Ensures accessible images

1. getByTitle - SVG titles, title attributes

- Rare, when other queries unavailable

1. getByTestId - Last resort only

- When no other query works - Not user-facing

Query Variants

Three variants for every query:

**getBy*** - Element must exist (throws if not found)

// ✅ Use when asserting element EXISTS
const button = screen.getByRole('button', { name: /submit/i });
expect(button).toBeDisabled();

**queryBy*** - Returns null if not found

// ✅ Use when asserting element DOESN'T exist
expect(screen.queryByRole('dialog')).not.toBeInTheDocument();

// ❌ WRONG - getBy throws, can't assert non-existence
expect(() => screen.getByRole('dialog')).toThrow(); // Ugly!

**findBy*** - Async, waits for element to appear

// ✅ Use when element appears after async operation
const message = await screen.findByText(/success/i);

Common Mistakes

❌ Using container.querySelector

const button = container.querySelector('.submit-button'); // DOM implementation detail

✅ CORRECT - Query by accessible role

const button = screen.getByRole('button', { name: /submit/i }); // User-facing

❌ Using getByTestId when role available

screen.getByTestId('submit-button'); // Not how users find button

✅ CORRECT - Query by role

screen.getByRole('button', { name: /submit/i }); // How screen readers find it

❌ Not using accessible names

screen.getByRole('button'); // Which button? Multiple on page!

✅ CORRECT - Specify accessible name

screen.getByRole('button', { name: /submit/i }); // Specific button

❌ Using getBy to assert non-existence

expect(() => screen.getByText(/error/i)).toThrow(); // Awkward

✅ CORRECT - Use queryBy

expect(screen.queryByText(/error/i)).not.toBeInTheDocument();

User Event Simulation

Always use userEvent over fireEvent for realistic interactions.

userEvent vs fireEvent

Why userEvent is superior:

• Simulates complete interaction sequence (hover → focus → click → blur)
• Triggers all associated events
• Respects browser timing and order
• Catches more bugs

// ❌ WRONG - fireEvent (incomplete simulation)
fireEvent.change(input, { target: { value: 'test' } });
fireEvent.click(button);

// ✅ CORRECT - userEvent (realistic simulation)
const user = userEvent.setup();
await user.type(input, 'test');
await user.click(button);

Only use fireEvent when:

• userEvent doesn't support the event (rare)
• Testing non-standard browser behavior

userEvent.setup() Pattern

Modern best practice (2025):

// ✅ CORRECT - Setup per test
it('should handle user input', async () => {
  const user = userEvent.setup(); // Fresh instance per test
  render('<input aria-label="Email" />');

  await user.type(screen.getByLabelText(/email/i), 'test@example.com');
});

// ❌ WRONG - Setup in beforeEach
let user;
beforeEach(() => {
  user = userEvent.setup(); // Shared state across tests
});

it('test 1', async () => {
  await user.click(...); // Might affect test 2
});

Why: Each test gets clean state, prevents test interdependence.

Common Interactions

Clicking:

const user = userEvent.setup();
await user.click(screen.getByRole('button', { name: /submit/i }));

Typing:

await user.type(screen.getByLabelText(/email/i), 'test@example.com');

Keyboard:

await user.keyboard('{Enter}'); // Press Enter
await user.keyboard('{Shift>}A{/Shift}'); // Shift+A

Selecting options:

await user.selectOptions(
  screen.getByLabelText(/country/i),
  'USA'
);

Clearing input:

await user.clear(screen.getByLabelText(/search/i));

Async Testing Patterns

UI frameworks are async by nature (state updates, API calls, suspense). Testing Library provides utilities for async scenarios.

findBy Queries

Built-in async queries (combines getBy + waitFor):

// ✅ CORRECT - Wait for element to appear
const message = await screen.findByText(/success/i);

// Under the hood: retries getByText until it succeeds or timeout

When to use:

• Element appears after async operation
• Loading states disappear
• API responses render content

Configuration:

// Default: 1000ms timeout
const message = await screen.findByText(/success/i);

// Custom timeout
const message = await screen.findByText(/success/i, {}, { timeout: 3000 });

waitFor Utility

For complex conditions that findBy can't handle:

// ✅ CORRECT - Complex assertion
await waitFor(() => {
  expect(screen.getByText(/loaded/i)).toBeInTheDocument();
});

// ✅ CORRECT - Multiple elements
await waitFor(() => {
  expect(screen.getAllByRole('listitem')).toHaveLength(10);
});

waitFor retries until:

• Assertion passes (doesn't throw)
• Timeout reached (default 1000ms)

Common mistakes:

❌ Side effects in waitFor

await waitFor(() => {
  fireEvent.click(button); // Side effect! Will click multiple times
  expect(result).toBe(true);
});

✅ CORRECT - Only assertions

fireEvent.click(button); // Outside waitFor
await waitFor(() => {
  expect(result).toBe(true); // Only assertion
});

❌ Multiple assertions

await waitFor(() => {
  expect(screen.getByText(/name/i)).toBeInTheDocument();
  expect(screen.getByText(/email/i)).toBeInTheDocument(); // Might not retry both
});

✅ CORRECT - Single assertion per waitFor

await waitFor(() => {
  expect(screen.getByText(/name/i)).toBeInTheDocument();
});
expect(screen.getByText(/email/i)).toBeInTheDocument();

❌ Wrapping findBy in waitFor

await waitFor(() => screen.findByText(/success/i)); // Redundant!

✅ CORRECT - findBy already waits

await screen.findByText(/success/i);

waitForElementToBeRemoved

For disappearance scenarios:

// ✅ CORRECT - Wait for loading spinner to disappear
await waitForElementToBeRemoved(() => screen.queryByText(/loading/i));

// ✅ CORRECT - Wait for modal to close
await waitForElementToBeRemoved(() => screen.queryByRole('dialog'));

Note: Must use queryBy* (returns null) not getBy* (throws).

Common Patterns

Loading states:

render('<div id="container"></div>');

// Simulate async data loading
const container = document.getElementById('container');
container.innerHTML = '<p>Loading...</p>';

// Initially loading
expect(screen.getByText(/loading/i)).toBeInTheDocument();

// Simulate data load
setTimeout(() => {
  container.innerHTML = '<p>John Doe</p>';
}, 100);

// Wait for data
await screen.findByText(/john doe/i);

// Loading gone
expect(screen.queryByText(/loading/i)).not.toBeInTheDocument();

API responses:

const user = userEvent.setup();
render(`
  <form>
    <label>Search: <input name="search" /></label>
    <button type="submit">Search</button>
    <ul id="results"></ul>
  </form>
`);

await user.type(screen.getByLabelText(/search/i), 'react');
await user.click(screen.getByRole('button', { name: /search/i }));

// Wait for results (after API response)
await waitFor(() => {
  expect(screen.getAllByRole('listitem')).toHaveLength(10);
});

Debounced inputs:

const user = userEvent.setup();
render(`
  <label>Search: <input id="search" /></label>
  <ul id="suggestions"></ul>
`);

await user.type(screen.getByLabelText(/search/i), 'react');

// Wait for debounced suggestions
await screen.findByText(/react testing library/i);

MSW Integration

Mock Service Worker for API-level mocking.

Why MSW

Network-level interception:

• Intercepts requests at network layer (not fetch/axios mocks)
• Same mocks work in tests, Storybook, development
• No client-specific mocking logic
• Tests real request logic

// ❌ WRONG - Mocking fetch implementation
vi.spyOn(global, 'fetch').mockResolvedValue({
  json: async () => ({ users: [...] }),
}); // Tight coupling, won't work in Storybook

// ✅ CORRECT - MSW intercepts at network level
// Works in tests, Storybook, dev server
http.get('/api/users', () => {
  return HttpResponse.json({ users: [...] });
});

setupServer Pattern

In test setup file:

// test-setup.ts
import { setupServer } from 'msw/node';
import { handlers } from './mocks/handlers';

export const server = setupServer(...handlers);

beforeAll(() => server.listen());
afterEach(() => server.resetHandlers());
afterAll(() => server.close());

In handlers file:

// mocks/handlers.ts
import { http, HttpResponse } from 'msw';

export const handlers = [
  http.get('/api/users', () => {
    return HttpResponse.json({
      users: [
        { id: 1, name: 'Alice' },
        { id: 2, name: 'Bob' },
      ],
    });
  }),
];

Per-Test Overrides

Override handlers for specific tests:

it('should handle API error', async () => {
  // Override for this test only
  server.use(
    http.get('/api/users', () => {
      return HttpResponse.json(
        { error: 'Server error' },
        { status: 500 }
      );
    })
  );

  render('<div id="user-list"></div>');

  // Simulate component fetching users
  fetch('/api/users').then(() => {
    document.getElementById('user-list').innerHTML =
      '<p>Failed to load users</p>';
  });

  await screen.findByText(/failed to load users/i);
});

After test, afterEach resets to default handlers.

Accessibility-First Testing

Why Accessible Queries

Three benefits:

1. Tests mirror real usage - Query like screen readers do
2. Improves app accessibility - Tests force accessible markup
3. Refactor-friendly - Coupled to user experience, not implementation

// ❌ WRONG - Implementation detail
screen.getByTestId('user-menu');

// ✅ CORRECT - Accessibility query
screen.getByRole('button', { name: /user menu/i });

If accessible query fails, your app has an accessibility issue.

ARIA Attributes

When to add ARIA:

✅ Custom components (where semantic HTML unavailable):

<div role="dialog" aria-label="Confirmation Dialog">
  <h2>Are you sure?</h2>
  ...
</div>

Query:

screen.getByRole('dialog', { name: /confirmation/i });

❌ DON'T add to semantic HTML (redundant):

<!-- ❌ WRONG - Semantic HTML already has role -->
<button role="button">Submit</button>

<!-- ✅ CORRECT - Semantic HTML is enough -->
<button>Submit</button>

Semantic HTML Priority

Always prefer semantic HTML over ARIA:

<!-- ❌ WRONG - Custom element + ARIA -->
<div role="button" onclick="handleClick()" tabindex="0">
  Submit
</div>

<!-- ✅ CORRECT - Semantic HTML -->
<button onclick="handleClick()">
  Submit
</button>

Semantic HTML provides:

• Built-in keyboard navigation
• Built-in focus management
• Built-in screen reader support
• Less code, more accessibility

Testing Library Anti-Patterns

1. Not using screen object

❌ WRONG - Query from render result

const { getByRole } = render('<button>Submit</button>');
const button = getByRole('button');

✅ CORRECT - Use screen

render('<button>Submit</button>');
const button = screen.getByRole('button');

Why: screen is consistent, no destructuring, better error messages.

2. Using querySelector

❌ WRONG - DOM implementation

const { container } = render('<button class="submit-btn">Submit</button>');
const button = container.querySelector('.submit-btn');

✅ CORRECT - Accessible query

render('<button>Submit</button>');
const button = screen.getByRole('button', { name: /submit/i });

3. Testing implementation details

❌ WRONG - Internal state

const component = new Component();
expect(component._internalState).toBe('value'); // Private implementation

✅ CORRECT - User-visible behavior

render('<div id="output"></div>');
expect(screen.getByText(/value/i)).toBeInTheDocument();

4. Not using jest-dom matchers

❌ WRONG - Manual assertions

expect(button.disabled).toBe(true);
expect(element.classList.contains('active')).toBe(true);

✅ CORRECT - jest-dom matchers

expect(button).toBeDisabled();
expect(element).toHaveClass('active');

Install: npm install -D @testing-library/jest-dom

5. Manual cleanup() calls

❌ WRONG - Manual cleanup

afterEach(() => {
  cleanup(); // Automatic in modern Testing Library!
});

✅ CORRECT - No cleanup needed

// Cleanup happens automatically

6. Wrong assertion methods

❌ WRONG - Property access

expect(input.value).toBe('test');
expect(checkbox.checked).toBe(true);

✅ CORRECT - jest-dom matchers

expect(input).toHaveValue('test');
expect(checkbox).toBeChecked();

7. beforeEach render pattern

❌ WRONG - Shared render in beforeEach

let button;
beforeEach(() => {
  render('<button>Submit</button>');
  button = screen.getByRole('button'); // Shared state
});

it('test 1', () => {
  // Uses shared button from beforeEach
});

✅ CORRECT - Factory function per test

const renderButton = () => {
  render('<button>Submit</button>');
  return {
    button: screen.getByRole('button'),
  };
};

it('test 1', () => {
  const { button } = renderButton(); // Fresh state
});

For factory patterns, see testing skill.

8. Multiple assertions in waitFor

❌ WRONG - Multiple assertions

await waitFor(() => {
  expect(screen.getByText(/name/i)).toBeInTheDocument();
  expect(screen.getByText(/email/i)).toBeInTheDocument();
});

✅ CORRECT - Single assertion per waitFor

await waitFor(() => {
  expect(screen.getByText(/name/i)).toBeInTheDocument();
});
expect(screen.getByText(/email/i)).toBeInTheDocument();

9. Side effects in waitFor

❌ WRONG - Mutation in callback

await waitFor(() => {
  fireEvent.click(button); // Clicks multiple times!
  expect(result).toBe(true);
});

✅ CORRECT - Side effects outside

fireEvent.click(button);
await waitFor(() => {
  expect(result).toBe(true);
});

10. Exact string matching

❌ WRONG - Fragile exact match

screen.getByText('Welcome, John Doe'); // Breaks on whitespace change

✅ CORRECT - Regex for flexibility

screen.getByText(/welcome.*john doe/i);

11. Wrong query variant for assertion

❌ WRONG - getBy for non-existence

expect(() => screen.getByText(/error/i)).toThrow();

✅ CORRECT - queryBy

expect(screen.queryByText(/error/i)).not.toBeInTheDocument();

12. Wrapping findBy in waitFor

❌ WRONG - Redundant

await waitFor(() => screen.findByText(/success/i));

✅ CORRECT - findBy already waits

await screen.findByText(/success/i);

13. Using testId when role available

❌ WRONG - testId

screen.getByTestId('submit-button');

✅ CORRECT - Role

screen.getByRole('button', { name: /submit/i });

14. Not installing ESLint plugins

Install these plugins:

npm install -D eslint-plugin-testing-library eslint-plugin-jest-dom

.eslintrc.js:

{
  extends: [
    'plugin:testing-library/dom', // For framework-agnostic
    // OR 'plugin:testing-library/react' for React
    'plugin:jest-dom/recommended',
  ],
}

Catches anti-patterns automatically.

Summary Checklist

Before merging UI tests, verify:

• Preferred: Using Vitest Browser Mode with real browser (not jsdom/happy-dom)
• All Playwright/Browser Mode tests are idempotent (no shared state between tests)
• Using getByRole as first choice for queries (built-in or Testing Library)
• Using expect.element() for auto-retrying assertions (Browser Mode)
• Using userEvent for interactions (CDP-based in Browser Mode, or @testing-library/user-event)
• Testing behavior users see, not implementation details
• No manual cleanup() calls (automatic)
• No manual act() calls (Browser Mode handles timing)
• MSW for API mocking (not fetch/axios mocks)
• Following TDD workflow (see tdd skill)
• Using test factories for data (see testing skill)
• For React-specific patterns (hooks, context, components), see react-testing skill