We want to fail the build if more console errors are introduced. For example, let's say console.error was called 30 times in the whole test suite. If another error is introduced this will increase to 31, which we don't want. Is there a way to prevent this?
For one test suite it is possible with:
const spy = jest.spyOn(console, "error");
let count = 0;
afterEach(() => {
count += spy.mock.calls.length;
});
afterAll(() => {
if (count > 2) {
throw Error(`oops error count: ${count}`);
}
});
but it would be nice to have this globally defined.
We solved this in a slightly different way:
// src/utils/testUtils
let consoleErrorSpy;
export const spyOnConsoleError = () => {
consoleErrorSpy = jest.spyOn(console, "error");
};
/**
* We are using this to prevent the console errors from increasing.
* These are our preferences in order of priority:
* 1. Don't call this method
* 2. Call this method at the end of a specific test (eg. for an error that can't be solved)
* 3. Call this method in `afterEach` (eg. for an async error that can't be solved)
*/
export const catchConsoleErrors = ({ silenced = [] } = {}) => {
const alwaysSilencedErrors = [
'<bug from a 3rd party library>'
];
const forbiddenCalls = [];
const silencedCalls = [];
for (const call of consoleErrorSpy.mock.calls) {
if (
new RegExp([...alwaysSilencedErrors, ...silenced].join("|")).test(call)
) {
silencedCalls.push(call);
} else {
forbiddenCalls.push(call);
}
}
for (const silencedCall of silencedCalls) {
// eslint-disable-next-line no-console
console.log("SILENCED\n---\n" + silencedCall.join(",") + "\n---");
}
expect(forbiddenCalls).toHaveLength(0);
// We clear the mock here so nothing happens if the method is called again for the same test,
// which is the case when this method is called in a specific test (file)
// as it is also called in `afterEach` in setUpTests.js
consoleErrorSpy.mockClear();
};
// some test file
afterEach(() => {
catchConsoleErrors({
silenced: [
"Warning: Can't perform a React state update on an unmounted component.*"
]
});
});
// src/setupTests.js
spyOnConsoleError();
afterEach(() => {
catchConsoleErrors();
});
Related
I'm trying to test a service that has a listener of the a custom Event Emitter in node with typescript and mocha, sinon.
My custom emmiter;
class PublishEmitter extends EventEmitter {
publish(id: string) {
this.emit('publish', id);
}
}
My service use case:
export default class PublishVehicle {
constructor(
private findVehicle: FindVehicle, // Service that contains find methods on repository
private updateVehicle: UpdateVehicle, // Service that contains update methods on repository
private logger: ILogger,
) {
this.producer = producer;
this.logger = logger;
}
listen() {
this.logger.log('debug', 'Creating listener on PublishEmitter');
this.publishListener = this.publishListener.bind(this);
pubsub.on('publish', this.publishListener);
}
/**
* Listener on PublishEmitter.
*
* #param event
*/
async publishListener(event: string) {
try {
const vehicle = await this.findVehicle.findById(event);
if (vehicle?.state === State.PENDING_PUBLISH) {
//
const input = { state: State.PUBLISH };
await this.updateVehicle.update(vehicle.id, input);
this.logger.log('debug', `Message sent at ${Date.now() - now} ms`);
}
this.logger.log('debug', `End Vehicle's Publish Event: ${event}`);
} catch (error) {
this.logger.log('error', {
message: `publishListener: ${event}`,
stackTrace: error,
});
}
}
}
and in my test file:
import chai from 'chai';
const { expect } = chai;
import sinon from 'sinon';
import { StubbedInstance, stubInterface } from 'ts-sinon';
import pubsub from './PublishEmitter';
describe('Use Case - Publish Vehicle', function () {
let mockRepository: MockVehicleRepository;
let publishVehicle: PublishVehicle;
let findVehicleUseCase: FindVehicle;
let updateVehicleUseCase: UpdateVehicle;
before(() => {
const logger = Logger.getInstance();
mockRepository = new MockVehicleRepository();
findVehicleUseCase = new FindVehicle(mockRepository, logger);
updateVehicleUseCase = new UpdateVehicle(mockRepository);
publishVehicle = new PublishVehicle(
findVehicleUseCase,
updateVehicleUseCase,
logger,
);
});
afterEach(() => {
// Restore the default sandbox here
sinon.restore();
});
it('Should emit event to publish vehicle', async () => {
const vehicle = { ... }; // dummy data
const stubFindById = sinon
.stub(mockRepository, 'findById')
.returns(Promise.resolve(vehicle));
const stubUpdate = sinon
.stub(mockRepository, 'update')
.returns(Promise.resolve(vehicle));
const spy = sinon.spy(publishVehicle, 'publishListener');
publishVehicle.listen();
pubsub.publish(vehicle.id);
expect(spy.calledOnce).to.be.true; // OK
expect(stubFindById.calledOnce).to.be.true; // Error (0 call)
expect(stubUpdate.calledOnce).to.be.true; // Error (0 call)
});
});
When I debug this test, indeed the methods are called but they seem to be executed after it has gone through the last expect lines.
The output:
1 failing
1) Use Case - Publish Vehicle
Should emit event to publish vehicle:
AssertionError: expected false to be true
+ expected - actual
-false
+true
UPDATE
Finally I was be able to solve my problem wrapping expect lines in setTimeout.
setTimeout(() => {
expect(spy.calledOnce).to.be.true; // OK
expect(stubFindById.calledOnce).to.be.true; // OK
expect(stubUpdate.calledOnce).to.be.true; // OK
done();
}, 0);
I would like to test following part of the code:
// ... code above
const created = async payload => {
const model = await db.collection('models').doc(payload.model)
.get() // <--- 1st .get() occurence
if (!model.exists) {
// Add product to the orphans collection
await db.collection('orphans').doc(payload.sku).set(payload)
} else {
// Grab the categories field
const categories = model.get('categories') // <--- 2nd .get() occurence
// Product is either empty or does not exists at all
if (!categories || categories.length < 1) {
// Add product to the orphans collection
await db.collection('orphans').doc(payload.sku).set(payload)
} else {
// Otherwise remove from the orphans collection
await deleted(payload.sku)
}
}
}
I do not know how to properly mock the file twice in the same callback. Here is what I get:
test.only('it should react when an event "created" has been fired', async () => {
const spy = jest.fn()
jest.doMock('#google-cloud/firestore', () => class {
collection () {
return {
doc: () => {
return {
get: () => {
return {
exists: () => {
spy()
}
}
},
set: () => {
spy()
}
}
}
}
}
})
const observer = require('./product')
await observer('created', {})
await expect(spy.mock.calls.length).toBe(1)
})
I get this error:
● it should react when an event "created" has been fired
TypeError: model.get is not a function
25 | } else {
26 | // Grab the categories field
> 27 | const categories = model.get('categories')
| ^
28 |
29 | // Product is either empty or does not exists at all
30 | if (!categories || categories.length < 1) {
at created (app/observers/product.js:27:30)
at Object.<anonymous>.module.exports (app/observers/product.js:6:28)
at Object.<anonymous> (app/observers/product.spec.js:34:3)
Test Suites: 1 failed, 1 total
Tests: 1 failed, 1 skipped, 2 total
Snapshots: 0 total
Time: 0.147 s, estimated 1 s
Ran all test suites matching /app\/observers\/product.spec.js/i.
What is the working solution to test two scenarios of the same mocked get() method ?
In your code :
const model = await db.collection('models').doc(payload.model)
.get() // <--- 1st .get() occurence
If we look at your mock, the get method of doc returns :
{
exists: () => {
spy()
}
}
There are no property named get, so it is undefined (and not a function).
I guess you just have to change this part to :
{
exists: true, // can be false
get: spy,
}
And your problem should be solved.
Btw, you can also change the mock of set method to set: spy. Or you can keep it to set: () => { spy() }, but you should at least return the value if you want to mock it : set: () => { spy() }.
Now, about how to properly mock multiple times, here's what you can do :
const observer = require('./product')
const spyGet = jest.fn()
const spySet = jest.fn() // I like having different mocks, if one function use get & set, tests will be clever & more readable if you use different spies
describe('on event "created" fired', () => {
const categories = []
beforeEach(() => {
// I put mocks here to make test more readable
jest.doMock('#google-cloud/firestore', () => class {
collection () {
return {
doc: () => {
return {
get: () => {
return {
exists: true,
get: spyGet,
}
},
set: spySet
}
}
}
}
})
spyGet.mockResolvedValueOnce(categories) // you can also use mockResolvedValue, but mockResolvedValueOnce allow you to mock with different values on the same test & same mock
})
it.only('should get categories', async () => {
await observer('created', {})
// here's all the ways you can test it
expect(spyGet).toBeCalledTimes(1)
expect(spyGet.mock.calls.length).toBe(1)
expect(spyGet).toBeCalledWith('categories')
expect(spyGet).toHaveBeenNthCalledWith(1, 'categories')
})
})
Note : You should reset & clear your mocks between tests manually (in a afterEach or beforeEach) if you don't set it into jest config.
I have this simple webworker which does some polling:
import { fetchLatestResults } from '../backend/fetchLatestResults';
let polling = false;
let symbolIds: string[] = [];
const fetchLatest = async () => {
if (!symbolIds.length) {
return;
}
try {
const results = await fetchLatestResults(symbolIds);
self.postMessage(results);
} catch (e) {
// noop for now
}
};
if (!polling) {
polling = true;
fetchLatest();
setInterval(fetchLatest, 10 * 1000);
}
self.onmessage = (e: MessageEvent) => {
symbolIds = e.data;
fetchLatest();
};
How can I unit test this with jest? There is not a lot of code, business logic is extracted, but still there are few cases that I'd like to test, like that fetchLatestResults is not invoked until symbolIds are set, etc. ?
I could probably extract onmessage logic into a function, export it, import it on tests, and call it, but what about postMessage ?
I have the following js function:
const modelUtils = {
modelingObj(obj, stringVal = ‘RE’) {
let newObj;
//my logic for setting value of newObj
return newObj;
};
export default modelUtils;
I want to test and see that based on a specific params I get a particular result, the issue is I’m always returning back an empty object.
Test.js
import modelUtils from '../modelUtils';
jest.unmock('../modelUtils');
describe(' testing modelUtils', () => {
let test;
const mockData = {
myProperty: [],
};
describe('testing modelingObj function', () => {
it('For my first test’, () => {
test = modelUtils.mockData, ‘TR’);
expect(test).toEqual({ myProperty: [] });
});
});
});
I have a method that logs a message via one function in a node environment and via a different function in a browser environment. To check whether I am in a node or browser environment I use the libraries detect-node and is-browser like so:
const isNode = require('detect-node');
const isBrowser = require('is-browser');
log(level, message, data) {
if (isNode) {
this.nodeTransport.log(level, this.name, message, data);
}
if (isBrowser) {
this.browserTransport.log(level, this.name, message, data);
}
}
The variables isNode and isBrowser are set to true and false (automatically via the package) depending on, well, if I'm in a browser or in a node env.
Now I want to test this behavior using jest so I need to mock these npm packages. This is what I tried:
function setup() {
const loggerName = 'Test Logger';
const logger = new Logger(loggerName);
logger.nodeTransport = { log: jest.fn() };
logger.browserTransport = { log: jest.fn() };
logger.splunkTransport = { log: jest.fn() };
return { logger, loggerName };
}
test('it should call the the appropriate transports in a node environment', () => {
const { logger } = setup();
const message = 'message';
jest.mock('detect-node', () => true);
jest.mock('is-browser', () => false);
logger.log('error', message, []);
expect(logger.nodeTransport.log).toHaveBeenCalled();
expect(logger.browserTransport.log).not.toHaveBeenCalled();
});
test('it should call the the appropriate transports in a browser environment', () => {
const { logger } = setup();
const message = 'message';
jest.mock('detect-node', () => false);
jest.mock('is-browser', () => true);
logger.log('error', message, []);
expect(logger.nodeTransport.log).not.toHaveBeenCalled();
expect(logger.browserTransport.log).toHaveBeenCalled();
});
You see, I am using jest.mock to mock detect-node and is-browser and give it different return values. However, this just does not work. The first test is green because (I assume) Jest runs in node, but the second test fails saying
Expected mock function not to be called but it was called with:
["error", "Test Logger", "message", []]
Use .mockClear() to reset the mock calls between tests.
afterEach(() => {
logger.nodeTransport.log.mockClear();
logger.browserTransport.log.mockClear();
});