playwright-expert

Create a class per page that exposes locators as readonly properties and methods for user actions, then instantiate it in tests with the page object. Use custom fixtures via test.extend() to share page objects and authenticated state across tests.

Use getByRole() for buttons, links, and inputs first, then getByLabel() for form fields and getByTestId() for non-semantic elements. Avoid CSS class selectors and XPath since they break easily when the UI changes.

Flakiness usually comes from race conditions, animations, network timing, or shared state between tests. Fix it by relying on auto-waiting locators, using waitForResponse for network calls, isolating test state, and never using waitForTimeout.

Use page.route() with route.fulfill() to return mock JSON responses, route.continue() to pass requests through, or route.fetch() to modify real responses. You can also record and replay responses with routeFromHAR().

Yes, playwright.config.ts supports projects for Chromium, Firefox, WebKit, and mobile device emulation like Pixel 5 or iPhone 13 using the devices descriptor. Tests run in parallel across all configured projects.

Run with PWDEBUG=1 to open the inspector, use --headed to watch the browser, or enable trace: 'retain-on-failure' and inspect the trace with npx playwright show-trace. The --ui flag opens an interactive UI mode for step-by-step debugging.