4eeaca7ab8
refs https://github.com/TryGhost/Team/issues/2371 - extracting the re-used actions to utils allows tests to be self-descriptive rather than relying on comments and keeps the selectors and related actions in one place to help refactoring if/when they change |
||
|---|---|---|
| .. | ||
| admin | ||
| portal | ||
| utils | ||
| README.md | ||
Browser testing
Install
As per the docs, run the following to install the supported browsers for Playwright:
npx playwright install
Running tests
Run the browser test suite with yarn test:browser.
Record test instructions
After installing PlayWright, start to record tests using yarn test:browser:record.
Available flags:
--admin- Runs a test starting in Ghost Admin--no-setup- When testing Ghost Admin, prevents the automated setup from running (for testing the setup wizard)--fixtures="posts,users"- Install a set of fixtures, given as a comma-delimited list
When the window loads, hit the record button and click around. All of the link click steps will need to be tidied up to use better selectors, and any page.goto calls should be dropped (as they are likely a result of clicking links). The test generator is a useful assistant, but be wary of taking anything it generates as correct.
Writing tests
Recording tests will allow you to execute the steps in a browser that a user would follow. It's important to add assertions to ensure that we're not just testing that the site doesn't crash, but that we see the expected values on the page.
The test suite uses beforeAll and afterAll to setup Ghost and install fixtures. Each set of fixtures should have a new describe block to start a new instance of Ghost to be tested against.