What is it ?

A minimal javascript test framework.

• Concise output by default
• Small self-evident codebase
• Command line interface with watch mode
• < 50 kB

How to use it

Create or navigate to a Node.js project directory

mkdir my-project
cd my-project
npm init -y

Install t3st (optional, but recommended)

npm i -d t3st

Generate a test

npx t3st gen hello

The above command creates a test file hello.js similar to this:

module.exports = async ({ test, throws, equal, check }) => {

    // const unit = require('...')

    return [
        test("description", () => {
            equal(2, 1 + 1)
            check(1, 1 + 1, (a, b) => a + b == 3)
        }),
        , throws("expects erros to be thrown", () => {
            throw 'uncomment this line to cause failing test'
        })
        , await test("async test has to be awaited.", async () => {
            equal({}, {})
        })]
}

Run the tests

npx t3st

Continuously re-run the tests when any code changes

npx t3st -w

Edit tests/hello.js to see output for failing tests.

Add the noisy -n flag to get more output

npx t3st -n

View command line documentation via the terminal

npx t3st help

Use with scripts / build servers:

t3st sets an exit code of 0 if all tests succeeded.

An exit code of 1 is set when:

• Any tests failed
• No tests are found
• Unhandled promise rejections were detected.

To prevent writing output to the console, use the silent option -s or --silent (just the error code is set).

To specify a different directory than $(pwd)/tests use the -d or --dir option.

Project clia is an example of how t3st can be used on a build server, with this CI/CD github workflow definition and this package.json config.

How it works

t3st recursively reads all javascript test files in a directory and imports them.

The default function that each test exports is called with the framework validation functions as arguments.

Invoked tests produce a collection of test results:

• Passing tests, objects with a single property: { description }
• Failing tests, objects with other additional properties, eg. { description, error, .. }

These results are used to build up a report to display for the user.

Future work

Stand alone tests aren't supported (yet?), a good alternative is here

Keeping the test framework small, yet feature rich-enough is a balancing act between scope-creep and simplicity.

If you miss a feature that you really need or find a bug, please reach out / send a PR.