tst v7.3.0

tst

Test without efforts.

• tape API
• no tooling, vanilla ESM
• async functions support
• inspectable errors
• correct stacktrace with sourcemaps
• good l&f in browser/node
• supports assert, chai etc.
• tiny bundle, 0dep

usage

import test, { ok, is, not, throws } from 'tst.js'

test('pass', () => {
	ok(true);
	ok(true, 'this time with an optional message');
	ok('not true, but truthy enough');

	is(1 + 1, 2);
	is(Math.max(1, 2, 3), 3);
	is({}, {})

	throws(() => {
		throw new Error('oh no!');
	}, /oh no!/);
})

test('fail', () => {
	is(42, '42');
	is({}, {x:1});
})

Creates output in console:

api

• test.skip − bypass test, mutes output
• test.only − run only the indicated test, can be multiple
• test.todo − bypass test, indicate WIP sign
• test.demo − demo run, skips failed assertions.

assert

• ok(a, msg?) − generic truthfulness assert
• is(a, b, msg?) − assert with Object.is for primitives and deepEqual for objects
• not(a, b, msg?) - assert with !Object.is for primitives and !deepEqual for objects
• any(a, [a, b, c], msg?) − assert with optional results
• almost(a, b, eps, msg?) − assert approximate value/array
• same(listA, listB, msg?) − assert same members of a list/set/map/object
• throws(fn, msg?) − fn must throw
• pass(msg), fail(msf) − pass or fail the whole test.

why?

Testing should not involve maintaining test runner. It should be simple as tap/tape, working in browser/node, ESM, with nice l&f, done in a straightforward way. I wasn't able to find such test runner that so I had to create one.

similar

• uvu
• tape-modern
• @goto-bus-stop/tape-modern
• brittle