Transforms Yadda features into working Testcafé tests.

First we need a feature-file, say ./e2e/features/demo.feature:

Feature: Some description

Scenario: Perform a single action

  Given one step
  When I run another step
  Then I test for something specific

Now we must define some steps to cover it, e.g. ./e2e/steps/demo.js:

import { Selector } from 'testcafe';

export default {
  'Given one step': () => async t => {
    await t
      .expect('body').ok();
  },

  'When I run another step': () => async t => {
    await t
      .click(Selector('button'));
  },

  'Then I test for $phrase': value => async t => {
    await t
      .expect(Selector('body').innerText)
      .contains(value);
  },
};

Finally we can generate the test-files and execute them:

$ bdd-tc e2e/features -- testcafe --color chrome:headless

Steps are labeled functions that receive arguments and return actual test functions.

Those calls are inlined on the generated tests, but its code is actually imported:

import $step0 from '../steps/demo.js';

fixture `Some description`;

test(`Perform a single action`, async t => {
  await $step0[`Given one step`]()(t);
  await $step0[`When I run another step`]()(t);
  await $step0[`Then I test for \$phrase`]("something specific")(t);
});

Before and after hooks for tests can be defined too.

They're are similar to step functions:

export default {
  before: {
    namedHook: () => async t => {
      // do something
    },
  },

  after: {
    // etc.
  },

  // use @path as input
  url(path = '/') {
    return process.env.BASE_URL + path;
  },
};

Now you can reference them with @before and @after annotations respectively:

@after=doSomething
Feature: Some description

@before=namedHook,somethingElse
Scenario: Perform a single action

  Given one step
  When I run another step
  Then I test for something specific

Depending on the context, beforeEach/afterEach or before/after is used automatically.

Additional $matchers can be defined within steps as follows:

export default {
  matchers: {
    test: '(foo|bar)',
    noMatch: '(?:[^\\s]*)',
  },

  'When ask for $test': test => async t => {
    console.log(test); // foo OR bar
  },

  'Then verify $noMatch': noMatch => async t => {
    console.log(noMatch); // undefined
  },
};

Captures made from matchers will be passed as arguments, non-matched placeholders will be captured as (.+?) and passed too.

Use (?:<PATTERN>) to omit captured values from matched placeholders.

Built-in annotations are:

• @xsnapshot — Unique for features, disables any @snapshot from scenarios below
• @snapshot — Unique for scenarios, it'll take snapshots after each step!
• @before — Setup before/beforeEach from features and scenarios
• @after — Setup after/afterEach from features and scenarios
• @only — Append .only on generated fixture/test calls
• @skip — Completely omit fixture/test from generated code
• @page — Optional pathame, used only if url() is setup
• @url — Append .page calls on generated fixture/test calls

Given @snapshost value is passed as takeSnapshot's selector option, so it can be an array, in which case will fallback until one selector matches/exists.

Any other annotation is keept as input-data and passed through invoked hooks.

Multiple values using [ ;,] as separator will be treated as arrays, e.g.

@media=foo,bar

Complex values can be passed as JSON values, e.g.

@arr=["foo", "bar"]
@obj={"baz": "buzz"}
@str="Other value, with commas, etc."

In order to assist you during writing steps, you can leverage on:

• takeSnapshot(...) — Calls the same method from testcafe-blink-diff
• useSelectors(obj) — Object containing Selector(...) definitions, can be nested
• useFixtures(obj) — Object containing any values as fixtures, can be nested
• getVal(key) — Validate and return value from registered fixtures, see above
• getEl(key) — Validate and return selector from registered ones, see above
• $(...) — Shortcut for Selector(...), same options as original call

Importing the bdd-tc/matchers module you gain access to:

• jsf(schema[, options]) — Generate one or many samples from given JSON-Schema¹
• faker[...] — Faker.js instance - see demo
• chance[...] — Chance.js instance - see docs
• gen([type[, schema]]) — Generate a sample based on any given type, additional JSON-Schema is applied if given
• date([step]) — Random Date object, given optional step: seconds, minutes, hours, days, months or years
• pick(dataset) — Pick any value from given dataset, even work with strings!
• oneOf(dataset, whereField, fieldValue) — Find any item on dataset that matches field/value
• number([min[, max]]) — Returns a random number within min/max boundaries
• randexp(regexp) — Return a string generated from any given RegExp
• shuffle(dataset) — Copy, randomize and returns any given dataset

¹ We're using json-schema-faker under the hood to generate those.

• npm install — Setup dependencies
• npm run e2e — Run defined e2e tests
• npm run test:ci — To run all unit-tests

Inspect the generated results from E2E snapshots:

• npm run report:ui
• open generated/index.html