Option 3: run test in Node, run code being tested in a headless browser.
This works best when the code being tested needs to check how things are rendered (e.g. CSS styles), simulating interactions (e.g. what happens when the user clicks a specific button) or any other runtime-specific behaviors (e.g. CORS restrictions).
Note that you can mix Option 2 and 3 together: you can still take advantage of skipping the browser in tests that don't need it, and use a headless browser for tests that need one.
The library Playwright provides an API for manipulating headless browsers in a similar way as you would manually with the Developer Console.
Unlike Option 1, this doesn't require a human to run the test, so other actions could be chained (such as releasing a new version of a library, or sending an alert that the build is brhey toastedoken).
The other options:
First, install Node.
Then install the project dependencies:
npm install
Then run the test script (it runs the command defined in package.json):
npm test
The file /.github/worksflows/tests.yml signals Github to re-run the tests
when new code is available, using Github Actions.
The results are displayed in the tab Actions of the repository.
Depending on your settings, you would also receive a notification with a direct link to the results after a test has run.
The .vscode folder is optional: it merely defines the path to the tests
and recommends a few matching extensions.
This way, you could run specific tests on their own in the debugger, using Tests Explorer, and get direct feedback in the editor.
