A test runner for smallish node projects with an opinionated take on mocking and very few dependencies. It helps you hunt for Bugs... get it?
Dependencies
Package | Why |
---|---|
pirates |
for patching require |
stack-trace |
for identifying call sites |
$ npm install elmer-fudd
package.json
src/
multiply.js
scale.js
test/
multiply.test.js
package.json
{
"elmer-fudd": {
"ext": "test.js",
"root": "test",
"alias": {
"@src": "src"
}
},
"scripts": {
"test": "elmer-fudd"
},
}
src/multiply.js
const scale = require('./scale');
module.exports = (value) => value * scale;
src/scale.js
module.exports = 10;
test/multiply.test.js
const { test, assert } = require('elmer-fudd');
test({
name: 'Multiply without mocking',
unit: '@src/multiply',
spec: (multiply) => {
assert.equal(multiply(5), 50);
}
});
test({
name: 'Multiply with mocked scale',
unit: '@src/multiply',
mock: [
['@src/scale', 2]
],
spec: (multiply) => {
assert.equal(multiply(5), 10);
}
});
test
takes a “test object” as an input, which allows you to specify a name, the unit you wish to test, any mocks you want to provide for dependencies, and a “spec” that runs assertions. Below are a few examples to get you started:
Test with just a spec
const { test, assert } = require('elmer-fudd');
test({
name: 'Test with just a spec',
spec: () => {
assert.ok(true);
}
});
Unit testing a module
const { test, assert } = require('elmer-fudd');
test({
name: 'Unit testing a module',
unit: './path/to/double.js'
spec: (double) => {
assert.equal(double(2), 4);
}
});
Unit testing a module with mocked dependencies
const { test, assert } = require('elmer-fudd');
test({
name: 'Unit testing a module',
unit: './path/to/unit.js',
mock: [
['.path', { fake: true }],
],
spec: (double) => {
assert.equal(double(2), 4);
}
});
Unit testing a module with a spec object
const { test, assert } = require('elmer-fudd');
test({
name: 'Using a spec object',
unit: './path/to/sum.js',
spec: [
{ given: [1, 2], expect: 3 },
{ given: [1, 2, 3], expect: 6 },
]
});
assert
wraps node’s core assert.strict
library so that failures can be grouped with tests. For detailed information see the docs. Here are a few examples to give you some ideas:
TODO examples that build...
mockFn
returns a mock function you can use in your specs. Here is how you might use a mock function.
const { mockFn } = require('elmer-fudd');
const fn = mockFn();
fn.returns(2);
fn(); // returns 2
It is not a comprehensive solution, but if you need something more robust there is no reason you cannot use something like sinon instead. Mock functions have the following methods and properties:
fn.returns(value)
makes the mock return a specific valuefn.implementation(fn)
adds an implementation to the mockfn.calledWith(...args)
returns true if the mock has been called with these argsfn.throws(err)
when the mock is called, this error is thrownfn.resolves(value)
the mock returns a promise that resolves this valuefn.rejects(err)
the mock returns a promise that rejects with this errfn.reset()
resets the mock functionfn.calls
a getter that returns all the callsfn.count
a getter that returns how many times the mock has been called
You are allowed to mock the unit itself, which pared with "spec objects" creates a neat way you can write tests first without annoying failures. Then when you want to try the real unit you can comment out the test.
const { test, mockFn } = require('elmer-fudd')
test({
name: 'test first?',
unit: './my-busted-fn',
mock: [
['./my-busted-fn', mockFn().will([
{ given: [2], output: 'foo' },
{ given: [2, { extra: true }], output: { bar: 2 } },
])]
],
spec: [
{ given: [2], output: 'foo' },
{ given: [2, { extra: true }], output: { bar: 2 } }
],
})