Acting on the page

Every test is made of two things: actions and checks. You have seen checks already. Now let's cover the actions.

In Selenium an action always has two steps: find the element, then act on it.

await driver.findElement(By.css('[data-testid="email-input"]')).sendKeys('demo@nimbus.app');
await driver.findElement(By.css('[data-testid="login-button"]')).click();

The actions that cover most user behaviour:

• sendKeys(text) types into a field. Important: it appends. If the field already has text, clear it first:

const field = await driver.findElement(By.css('[data-testid="qty"]'));
await field.clear();
await field.sendKeys('3');

• click() clicks. The element must be visible and enabled at that moment, Selenium does not wait for that by itself. If the element appears after a delay, wait first (driver.wait(until.elementLocated(...))), then act.
• Dropdowns: for a real <select>, the Select helper class picks options by visible text or value. For custom dropdowns (divs styled as menus), you click them like any other elements.
• sendKeys(Key.ENTER) submits many forms, const { Key } = require('selenium-webdriver').

The discipline Selenium teaches, which the other tools hide: know what state the page is in before you act. A click on an element that is not there yet throws NoSuchElementException; a click on a covered element fails. When an action fails, first ask "was the page ready?", and if not, wait for the condition that makes it ready, never for a fixed time.

The full interactions reference: Interacting with web elements.

The problem below is pure actions-and-assertions: remove the last item from the cart and prove the empty state appears.