@netflix/x-test

x-test

a simple, tap-compliant test runner for the browser

Spec

• importable as type="module"
• is interoperable with TAP Version 13
• nested sub-tests run in an iframe
• has a recognizable testing interface

Interface

Testing

The following are exposed in the testing interface:

• test: Creates a sub-test in an iframe based on given src html page.
• it: The smallest testing unit--asynchronous.
• skip: An it whose callback is not run and which will pass.
• todo: An it whose callback is run and is expected to fail.
• waitFor: Ensures the test does not exit until given promise settles.
• assert: Simple assertion call that expects a boolean.

Events

• x-test-ping: root responds ('x-test-pong', { status: 'started'|'ended' })
• x-test-ended: all tests have completed or we bailed out
• x-test-bail: [internal] signal to quit test early
• x-test-queue: [internal] queues a new test
• x-test-next: [internal] destroy current test and create a new one
• x-test-it-started: [internal] user defined a new "it"
• x-test-it-ended: [internal] user-defined "it" completed

Parameters

The following parameters can be passed in via a url search:

• x-test-no-reporter: turns off custom reporting tool ui

Execution

Both test and it calls will execute in order**. test calls will boot the given html page in an iframe. Such iframes are run one-at-a-time. All invoked it calls await the completion of previously-invoked it calls.

**This is not the case if you nest it--but that's an anti-pattern anyhow.

Usage with puppeteer

See test.js for an example of how you can use puppeteer to run your app's tests and log the resulting TAP output to the console.

Parsing TAP

There are many TAP formatters. Importantly, as long as you can pipe the TAP output to another program, the output should be interoperable.