Test Lifecycle Events

There are a few moments in the life of a test where you might want to attach some setup or teardown logic. Node-tap implements these using the following interfaces.

CLI: --before=before-tests.js, --after=after-tests.js

To run a script before any tests are executed, specify it as a --before config value. To run a script after all tests are done executing, specify it as an --after config value.

These can be set either on the CLI, in a .taprc file, or package.json file. (See Configuring Tap.)

--before and --after file's output will be sent to the parent's terminal, and outside of any reporters. It's generally not a great idea to have them output TAP, since that can cause the test run to generate invalid output.

If the script exits in error (either via a status code or being killed by a signal), then the test run will be aborted and exit in error. An --after script will run even if the test run bails out.

A defined --before or --after script will be omitted if it would have been included as a test file. So, it's fine to do something like tap --before=test/setup.js --after=test/teardown.js test/*.js.

There is no provided way to communicate context from a --before or --after program, since they run in separate processes from the test scripts, but since they are guaranteed to be run before or after any parallel testing, it is safe to have them write data to files that can be read by test scripts.

The other functions referenced below are for use within a test program.

t.beforeEach(fn(done, childTest))

Before any child test (or any children of any child tests, etc.) the supplied function is called with two arguments. The first is a callback to indicate that the function is complete. The second is the test object that it's prefixing.

If the function returns a Promise, then that is used as the indication of doneness. Thus, async functions automatically end when all of their awaited Promises are complete.

t.afterEach(fn(done, childTest))

This is called after each child test (or any children of any child tests, on down the tree). Like beforeEach, it's called with a done callback as the first argument, and the child test object as the second, and can return a Promise.

t.teardown(fn())

When the test is completely finished, the teardown functions are called. They do not receive a done callback, but may return a Promise to perform asynchronous actions.

Why no t.before()?

There is no t.before() because there are other options for doing this:

• If the actions are synchronous, you can put them before your tests.
• If the actions are asynchronous, you can use a t.test().

t.on('end')

The end event fires when the test is completely finished, and all of its teardown functions have completed.

This is just a normal EventEmitter event, so it doesn't support any sort of async actions.

t.context

You can use the t.context object to track details specific to a test. For example, a beforeEach function might create a database connection, and then an afterEach function might shut it down cleanly.

const myDataBase = require('my-special-db-thingie')

t.beforeEach((done, t) => {
  t.context.connection = myDataBase.connect()
  done()
})

t.afterEach((done, t) => {
  t.context.connection.disconnect()
  done()
})

t.test('read and write', t => {
  const conn = t.context.connection
  conn.write('foo', 'bar')
  t.equal(conn.read('foo'), 'bar')
  t.end()
})