I can't seem to get the before Hook to be asynchronous.
I've used a before hook to start it up, which it does, but I'm trying to pick up when it fails, so that I can fail/skip the whole block (since there is no server to test with).
I've used the [async] before hook all throughout my code, but in this case it's not waiting for my done to be fired.
Basically, got a bunch of MochaJS tests, and for one block I need to start up a Web Server so that I can make Requests to it to test the RESTful API. I want to make sure that it started successfully, hence the async method.
describe('RESTful Tests', function(){
// Start up the whole express server for these API tests
before(function(done) {
try {
server.on('error', (e) => {
console.error("HERE");
done();
});
listener = server.listen(port, function() {
console.log("Server Started");
done();
});
console.error("AFTER LISTEN");
} catch(err) {
console.error("ON Caught");
done();
}
console.error("At End!!");
});
This runs, and shows as an 1) "before all" hook in the test, but it does not wait for the done to be called. I'm only getting the output...
RESTful Tests
AFTER LISTEN
At End!!
1) "before all" hook
None of the other items show up, but I'll get an exception (expected as I've blocked the port) about the EADDRINUSE. I'm really struggling to find a way to catch this.
I understand that it's in a different (kinda) "thread", so I wouldn't pick up the error, but still.... why is the asynchronous done method not being obeyed?
Seems some of the documentation out there is a bit misleading...
The Node/ExpressJS documentation state to put the .on('error' method on the server object. Seems in this case it needed to be on the listener object!
So my final [working] code looks like this...
listener = server.listen(port, function() {
done();
}).on('error', (function(e) {
this.skip(e.code);
done();
}).bind(this));
Yay!!! :)
Related
This question already has answers here:
mocha with nodejs assert hangs/timeouts for assert(false) instead of error
(3 answers)
Closed 5 years ago.
I'm trying to unit test a controller method with mocha + chai + sinon, but I'm struggling to understand how the assertion works with these async methods. Here is my test:
it('should fetch all registered books and send them as json', (done) => {
res.json = sandbox.spy();
const books = [];
BookMock.expects('find').withArgs()
.resolves(books);
booksController.index(req, res, () => {
res.json.should.have.been.calledWith(200, { data: books });
done();
});
});
The problem with this test is that if I change the assertion to expect a 300 instead of 200 parameter, for example, my test will stop (fail) and never call done(), leading to a generic failure which tells nothing more than a 'Test failed because done was never called', which says nothing related to the reason the test failed at all.
If I keep the test like that and my controller method does everything good, the test passes and it's all ok, but if the controller does anything wrong, the test fails and doesn't tell me the reason it failed (which should be "res.json was not called with bla bla bla parameters").
I am also not sure this is the correct way to unit test controller methods, since they don't return a Promise I can't use chai-as-promised or promise chains at all, so I used the next() callback (which always gets called using restify) to make the assertion, that might not be right, so I'm open to a full refactor to make this test the most correct.
Thank you in advance!
Digging a little bit more I've found this answer in other question: https://stackoverflow.com/a/26793989/4233017
So I updated my code and wrapped the assertion in a try/catch, ending up like this:
it('should fetch all registered books and send them as json', (done) => {
res.json = sandbox.spy();
const books = [];
BookMock.expects('find').withArgs()
.resolves(books);
booksController.index(req, res, () => {
try {
res.json.should.have.been.calledWith(200, { data: books });
done();
} catch (e) {
done(e);
}
});
});
Now when the test fails it gives me the correct error message. I could also do it with booleans as the answer says, but I think it's better this way.
Any other solution is still very much appreciated.
Ok, my mocha tests will pass if I comment out the 'before' and 'after' methods. I am sure that both of my errors are related to each other.
The 'after' method fails stating app.close isn't a function. The 'before' method fails saying it cant find 'app' on my line 7 (clearing server cache).
I am completely out of options or ideas. I would like to be able to start and stop my server at my command. This is the first time that I have attempted to include any type of 'before/after' methods to my mocha testing. working code below, but with my failing portion commented out. Any suggestions??
var request = require('supertest');
var app = require('../../server');
describe('server', function() {
before(function () {
//var app = require('../../server')();
//delete require.cache[require.resolve('app')];
});
after(function () {
//app.close();
});
describe('basic comms', function() {
it('responds to root route', function testSlash(done) {
request(app)
.get('/')
.expect('Content-type', /json/)
//.expect(res.message).to.equal('Hello World!')
.expect(200, done);
});
it('404 everything else', function testPath(done) {
//console.log('testing 404 response');
request(app)
.get('/foo/bar')
.expect(404, done);
});
});
});
In before you require your app in a different way than in line 2. Why would you not use already required app?
Example:
before(function () {
// here you can use app from line 2
});
Regarding app.close, where did you find this function?
Check Express docs:
http://expressjs.com/en/4x/api.html#app
To close express server, you can use this approach:
how to properly close node-express server?
Since start(), stop() will be removed in Qunit 2.0, what is the alternative for async setups and teardowns via the beforeEach, afterEach methods? For instance, if I want the beforeEach to wait for a promise to be finished?
QUnit basically wants people to stop using the global methods (not just start() and stop(), but also test(), expect(), etc). So, as of version 1.16.0, you should always use either the global namespace (QUnit) or the assert API argument passed into the test() functions. This includes the new async control:
QUnit.test( "testing async action", function( assert ) { // <-- note the `assert` argument here
var done = assert.async(); // tell QUnit we're doing async actions and
// hold onto the function it returns for later
setTimeout(function() { // do some async stuff
assert.ok( true, "This happened 100 ms later!" );
done(); // using the function returned from `assert.async()` we
// tell QUnit we're don with async actions
}, 100);
});
If you are familiar with the old start() and stop() way of doing things, you should see that this is extremely similar, but more compartmentalized and extensible.
Because the async() method call is on the assert argument into the test, it cannot be used in the beforeEach() function. If you have an example of how you were doing that before, please post it and we can try to figure out how to git it into the new way.
UPDATE
My mistake previously, the assert object is being passed into the beforeEach and afterEach callbacks on modules, so you should be able to do the same logic that you would do for a test:
QUnit.module('set of tests', {
beforeEach: function(assert) {
var done = assert.async();
doSomethingAsync(function() {
done(); // tell QUnit you're good to go.
});
}
});
(tested in QUnit 1.17.1)
Seeing that nobody has answered the beforeEach/afterEach part: a test suite is supposed to run as soon as the page loads. When that is not immediately possible, then resort to configuring QUnit:
QUnit.config.autostart = false;
and continue with setting up your test suite (initializing tests, feeding them to QUnit, asynchronously waiting for some components to load, be it AJAX or anything else), your site, and finally, when it's ready, just run:
QUnit.start();
QUnit's docsite has it covered.
Ember Qunit, has once exists beforeEach/setup, afterEach/teardown co-exist for a little while.
See PR: https://github.com/emberjs/ember-qunit/pull/125
I'm using promises to wrap asynchronous (Mongo) DB ops at the end of an (expressJS) route.
I want to try and figure out how to test the following code.
userService
userService.findOne = function (id) {
var deferred = q.defer();
User.findOne({"_id" : id})
.exec(function (error, user) {
if (error) {
deferred.reject(error);
} else {
deferred.resolve(user);
}
});
return deferred.promise;
};
userRoute
var user = function (req, res) {
var userId = req.params.id
, userService = req.load("userService");
// custom middleware that enables me to inject mocks
return userService.findOne(id)
.then(function (user) {
console.log("called then");
res.json({
msg: "foo"
});
}).catch(function (error) {
console.log("called catch");
res.json({
error: error
});
}).done();
};
Here's an attempt to test the above with mocha
userTest
it("when resolved", function (done) {
var jsonSpy = sinon.spy(httpMock.res, "json")
, httpMock = require("/path/to/mock/http/object")
, serviceMock = require("/path/to/mock/service"),
, deferred = q.defer()
, findStub = sinon.stub(serviceMock, "findOne")
.returns(deferred.promise)
, loadStub = sinon.stub(httpMock.req, "load")
.returns(serviceMock),
retPromise;
// trigger route
routes.user(httpMock.req, httpMock.res);
// force promise to resolve?
deferred.resolve();
expect(jsonSpy.called).to.be.true; // fails
// chai as promised
retPromise = findStub.returnValues[0];
expect(retPromise).to.be.fulfilled; // passes
});
the http mock is just an empty object with no-ops where expressJS would normally start rendering stuff. I've added some logging inside those no-ops to get an idea on how this is hanging together.
This isn't really working out. I want to verify how the whole is integrated, to establish some sort of regression suite - but I've effectively mocked it to smithereens and I'm just testing my mocks (not entirely successfully at that).
I'm also noticing that the console logs inside my http mocks triggered by then and catch are firing twice - but the jsonSpy that is invoked inside the actual code (verified by logging out the sinon spy within the userRoute code) is not called in test.
Has anyone got some advice on integration testing strategies for express apps backed by Mongo?
It looks to me like you're not giving your promise an opportunity to fire before you check if the result has been called. You need to wait asynchronously for userService.findOne()'s promise chain to complete before jsonSpy.called will be set. Try this instead:
// start of code as normal
q.when(
routes.user(httpMock.req, httpMock.res),
function() { expect(jsonSpy.called).to.be.true; }
);
deferred.resolve();
// rest of code as normal
That should chain off the routes.user() promise and pass as expected.
One word of caution: I'm not familiar with your framework, so I don't know if it will wait patiently for all async events to go off. If it's giving you problems calling back into your defer chain, you may want to try nodeunit instead, which handles async tests very well (IMO).
I try to create e2e tests with karma and jasmine with yeoman. In my karma-e2e.conf.js I add jasmine:
files = [
JASMINE,
JASMINE_ADAPTER,
ANGULAR_SCENARIO,
ANGULAR_SCENARIO_ADAPTER,
'test/e2e/**/*.js'
];
A need async testing so I need to use runs, waits, waitsFor (https://github.com/pivotal/jasmine/wiki/Asynchronous-specs)
But if I try to use it:
it('test', function () {
runs(function () {
...
});
});
Scenatio test runner returns this:
TypeError: Cannot call method 'runs' of null
at runs (http://localhost:8080/adapter/lib/jasmine.js:562:32)
at Object.<anonymous> (http://localhost:8080/base/test/e2e/eduUser.js:42:3)
at Object.angular.scenario.SpecRunner.run (http://localhost:8080/adapter/lib/angular-scenario.js:27057:15)
at Object.run (http://localhost:8080/adapter/lib/angular-scenario.js:10169:18)
I don't know where the problem is. Can you help me please?
Angular e2e tests with Karma don't and can't use the JASMINE adapter. Instead you have the ANGULAR_SCENARIO_ADAPTER which has a similar feel to writing Jasmine tests.
All commands in the adapter's API are asynchronous anyway. For example element('#nav-items').count() doesn't return a number, it returns a Future object. Future objects are placed in a queue and executed asynchronously as the runner progresses. To quote the API docs:
expect(future).{matcher}:
[...] All API statements return a future object, which get a value assigned after they are executed.
If you need to run your own asynchronous test code, you can extend the adapter's DSL, this is easier than it might sound. The idea is that you return your own Future which can be evaluated by a matcher such as toBe(). There are some examples on how to do this in the e2e-tests.js Gist from Vojta. Just remember to call done(null, myRetrunValue); when your test code is successful (myRetrunValue is the value evaluated by your matcher). Or done('Your own error message'); if you want the test to fail.
UPDATE: In response to question below. To simulate a login, first add a function called login to the dsl:
angular.scenario.dsl('login', function() {
return function(selector) {
// #param {DOMWindow} appWindow The window object of the iframe (the application)
// #param {jQuery} $document jQuery wrapped document of the application
// #param {function(error, value)} done Callback that should be called when done
// (will basically call the next item in the queuue)
return this.addFutureAction('Logging in', function(appWindow, $document, done) {
// You can do normal jQuery/jqLite stuff here on $document, just call done() when your asynchronous tasks have completed
// Create some kind of listener to handle when your login is complete
$document.one('loginComplete', function(e){
done(null, true);
}).one('loginError', function(e){
done('Login error', false);
});
// Simulate the button click
var loginButton = $document.find(selector || 'button.login');
loginButton.click();
})
};
});
And then call:
beforeEach( function()
{
expect( login('button.login') ).toBeTruthy();
});