So I have a folder full of scripts that all resemble a structure like this
// Adapter-100.js
angular.module('myModule', ['myParentFactory', function(myParentFactory) {
return angular.extend(myParentFactory, {
"someFunctionA" : function() {},
"someFunctionB" : function() {},
"someFunctionC" : function() {}
});
}]);
And my test just checks that they have these three methods, trouble is there is about 100 of these files (they're adapters for communicating with a server)
Here is a representation of my tests file
// api-adapter-tests.js
describe('Unit: EndPointMethods', function() {
var values, factory, adapter;
// Boot the module
beforeEach(function() {
module('myModule');
inject(function ($injector) {
values = $injector.get('AppConsts');
factory = $injector.get('EndPointConnection');
adapter = $injector.get('TestAdapter'); // This needs to change to match what adapter is being tested
});
});
// Run some tests
describe('AppConsts', function() {
it('Should have an api_host key', function() {
expect(values).toBeDefined();
expect(values.api_host).toBeDefined();
expect(typeof values.api_host).toBe('string');
});
});
// Is this able to be generated to test each adapter independently?
describe('TestEndPointMethod has minimum functional definitions', function() {
it('should have 3 defined functions', function() {
expect(factory.consumeResponse).toBeDefined();
expect(factory.getEndPoint).toBeDefined();
expect(factory.mutator).toBeDefined();
});
});
});
I don't want to have to write a separate describes/it block for each adapter but rather have Karma loop over all of these and create the tests on the fly (the tests are very unlikely to ever change)
I've Googled around for a solution to this but can't seem to find anything that does this kind of thing for me.
You can wrap your suites in a clojure and pass the Adapter you want to test: mocha will take care of running it in the right way - and so Karma.
function runSuiteFor(newAdapter){
return function(){
// api-adapter-tests.js
describe('Unit: EndPointMethods', function() {
var values, factory, adapter;
// Boot the module
beforeEach(function() {
module('myModule');
inject(function ($injector) {
values = $injector.get('AppConsts');
factory = $injector.get('EndPointConnection');
adapter = $injector.get(newAdapter); // set the Adapter here
});
});
// Run some tests
describe('AppConsts', function() {
it('Should have an api_host key', function() {
expect(values).toBeDefined();
expect(values.api_host).toBeDefined();
expect(typeof values.api_host).toBe('string');
});
});
// Is this able to be generated to test each adapter independently?
describe('TestEndPointMethod has minimum functional definitions', function() {
it('should have 3 defined functions', function() {
expect(factory.consumeResponse).toBeDefined();
expect(factory.getEndPoint).toBeDefined();
expect(factory.mutator).toBeDefined();
});
});
});
}
}
var adapters = ['MyTestAdapter1', 'MyTestAdapter2', etc...];
for( var i=0; i<adapters.length; i++){
runSuiteFor(adapters[i])();
}
Note: IE8 has some issues with this approach sometimes, so in case you're with Angular 1.2 bare in mind this.
Related
I have a javascript function which is supposed to behave differently when offline than online as a safeguard. I would like to have a Jasmine unit test which tests the function in both offline and online modes - e.g.,
// offline
describe('When there is no connection to the internet', function() {
beforeEach(function(){
spyOn(navigator, 'onLine').and.returnValue(false);
});
it('offline behavior happens', function() {
myFunction();
expect(something).not.toHaveBeenCalled();
});
});
// online
describe('When there is a connection to the internet', function() {
beforeEach(function(){
spyOn(navigator, 'onLine').and.returnValue(true);
});
it('online behavior happens', function() {
myFunction();
expect(something).toHaveBeenCalled();
});
});
However, I am unable to fake the value of navigator.onLine. In my before, I also tried:
navigator = {
'onLine': false
}
This didn't work either. To be thorough, I tried the same techniques above with window.navigator.onLine and that also didn't work.
Does anyone know how to mock being offline for a Jasmine test?
You have no chance to overwrite some built in properties. Though with Object.defineProperty() you can reconfigure builtin properties.
Jasmine 2.6 and newer
Since Jasmine 2.6 (release notes) there is spyOnProperty().
This is how you would spy navigator.onLine:
beforeEach(function(){
spyOnProperty(Navigator.prototype, 'onLine').and.returnValue(false);
});
Note that the prototype object Navigator is referenced, not it's instance window.navigator.
Jasmine pre 2.6
Build a facade for browser objects
With pre Jasmine 2.6 you can't spy attributes (or properties) directly.
I'd suggest to create a facade for such browser builtins using getter methods. Then you can mock those methods in your tests to return what you like.
const Browser = (function() {
return {
isOnline: function() {
return navigator.onLine;
}
};
})();
In your code:
if (Browser.isOnline()) {
// ...
}
In your test:
beforeEach(function(){
spyOn(Browser, 'isOnline').and.returnValue(false);
});
Do what spyOnProperty() does on your own
If you cannot upgrade from a pre 2.6 version for some reason, you could even spy manually by utilizing Object.defineProperty() and Object.getOwnPropertyDescriptor().
var onlineState,
origOnLineProp;
beforeEach(function() {
// Remember original config
origOnLineProp = Object.getOwnPropertyDescriptor(Navigator.prototype, "onLine");
onlineState = true;
// New behavior
Object.defineProperty(Navigator.prototype, "onLine", {
enumerable: origOnLineProp.enumerable,
configurable: origOnLineProp.configurable,
get: function() { return onlineState }
});
});
it("...", function() {
onlineState = false;
expect("code").toBe("correctly functioning in offline mode");
});
afterEach(function() {
// Restore original behavior
Object.defineProperty(Navigator.prototype, "onLine", {
enumerable: origOnLineProp.enumerable,
configurable: origOnLineProp.configurable,
get: origOnLineProp.get
});
});
You could even implement your own spyOnProperty() backport if you're desperate (that's beyond this answer though).
I had a similar issue, but in addition was trying to detect the situation when the browser went from an offline state to an online state, in order to try reconnecting an EventSource.
The code I was trying to test was this:
private startOfflineWatcher(): void {
const that = this;
window.addEventListener('offline', () => {
window.addEventListener('online', () => {
that.reconnectTimeout();
});
});
}
Using spyOnProperty(Navigator.prototype, 'onLine').and.returnValue(false); did not work in my case, because I had to simulate the state change. However, I was able to achieve this by simply dispatching two new events to window. My working Jasmine test looks like this now:
it('should start reconnect when getting online after being offline', () => {
spyOn<any>(service, 'reconnectTimeout');
const offlineEvent = new Event('offline');
const onlineEvent = new Event('online');
(service as any).startOfflineWatcher();
window.dispatchEvent(offlineEvent);
window.dispatchEvent(onlineEvent);
expect((service as any).reconnectTimeout).toHaveBeenCalled();
});
Depending on the use case, this might be more suitable than mocking the navigator.
This worked for me
window.navigator['__defineGetter__']('onLine', () => false); // false = OFFLINE
window.navigator['__defineGetter__']('onLine', () => true); // true = ONLINE
(I asked this question recently and accepted an answer but it's still not what I need.) I really need to create dynamic tests from data loaded from a module. Each item from the array will have it's own describe statement with certain protractor actions. My previous post has an answer that says to use an it statement, but I can't do that because there's too much going on.
My main problem is that the data doesn't get loaded in time for the describe. I had another suggestion to use VCR.js or something similar but I don't think those will work because I'm using a module. Is there a way I can save the data to a separate file and load it in? Would that be a good way to go?
var data = require('get-data'); //custom module here
describe('Test', function() {
var itemsArr;
beforeAll(function(done) {
data.get(function(err, result) {
itemsArr = result; //load data from module
done();
});
})
//error: Cannot read property 'forEach' of undefined
describe('check each item', function() {
itemsArr.forEach(function(item) {
checkItem(item);
});
});
function checkItem (item) {
var itemName = item.name;
describe(itemName, function() {
console.log('describe');
it('should work', function() {
console.log('it');
expect(true).toBeTruthy();
});
});
}
});
UPDATE:
I used Eugene's answer and came up with this. I can't test each individual study how I want because the it statement doesn't fire. Is this problem even solvable??
describe('check each item', function () {
it('should load data', function (done) {
browser.wait(itemsPromise, 5000);
itemsPromise.then(function(itemsArr) {
expect(itemsArr).toBeTruthy();
studyArr = itemsArr.filter(function (item) {
return item.enabled && _.contains(item.tags, 'study');
});
studyCount = studyArr.length;
expect(studies.count()).toEqual(studyCount);
checkItems(studyArr);
done();
});
});
function checkItems (itemsArr) {
itemsArr.forEach(function (item) {
describe(item.id, function () {
console.log('checkItems', item.id);
// doesn't work
it('should work', function (done) {
expect(false).toBeTruthy();
done();
});
});
});
}
});
You're trying to do something that Jasmine does not allow: generating tests after the test suite has started. See this comment on an issue of Jasmine:
Jasmine doesn't support adding specs once the suite has started running. Usually, when I've needed to do this, I've been able to know the list of options ahead of time and just loop through them to make the it calls. [...]
("adding specs" === "adding tests")
The point is that you can generate tests dynamically but only before the test suite has started executing tests. One corollary of this is that the test generation cannot be asynchronous.
Your second attempt does not work because it is trying to add tests to a suite that is already running.
Your first attempt is closer to what you need but it does not work either because describe calls its callback immediately, so beforeAll has not run by the time your describe tries to generate the tests.
Solutions
It all boils down to computing the value of itemsArr before the test suite start executing tests.
You could create a .getSync method that would return results synchronously. Your code would then be something like:
var data = require('get-data'); //custom module here
var itemsArr = data.getSync();
describe('Test', function() {
describe('check each item', function() {
itemsArr.forEach(function(item) {
checkItem(item);
});
});
[...]
If writing .getSync function is not possible, you could have an external process be responsible for producing a JSON output that you could then deserialize into itemsArr. You'd execute this external process with one of the ...Sync functions of child_process.
Here's an example of how the 2nd option could work. I've created a get-data.js file with the following code which uses setTimeout to simulate an asynchronous operation:
var Promise = require("bluebird"); // Bluebird is a promise library.
var get = exports.get = function () {
return new Promise(function (resolve, reject) {
var itemsArr = [
{
name: "one",
param: "2"
},
{
name: "two",
param: "2"
}
];
setTimeout(function () {
resolve(itemsArr);
}, 1000);
});
};
// This is what we run when were are running this module as a "script" instead
// of a "module".
function main() {
get().then(function (itemsArr) {
console.log(JSON.stringify(itemsArr));
});
};
// Check whether we are a script or a module...
if (require.main === module) {
main();
}
Then, inside the spec file:
var child_process = require('child_process');
var itemsArr = JSON.parse(child_process.execFileSync(
"/usr/bin/node", ["get-data.js"]));
describe('Test', function() {
itemsArr.forEach(function(item) {
checkItem(item);
});
function checkItem (item) {
var itemName = item.name;
describe(itemName, function() {
console.log('describe');
it('should work', function() {
console.log('it');
expect(true).toBeTruthy();
});
});
}
});
I've tested the code above using jasmine-node. And the following file structure:
.
├── data.js
├── get-data.js
└── test
└── foo.spec.js
./node_modules has bluebird and jasmine-node in it. This is what I get:
$ ./node_modules/.bin/jasmine-node --verbose test
describe
describe
it
it
Test - 5 ms
one - 4 ms
should work - 4 ms
two - 1 ms
should work - 1 ms
Finished in 0.007 seconds
2 tests, 2 assertions, 0 failures, 0 skipped
Try to use a promise, something like:
var deferred = protractor.promise.defer();
var itemsPromise = deferred.promise;
beforeAll(function() {
data.get(function(err, result) {
deferred.fulfill(result);
});
})
And then:
describe('check each item', function() {
itemsPromise.then(function(itemsArr) {
itemsArr.forEach(function(item) {
checkItem(item);
});
});
});
Another solution I can think of is to use browser.wait to wait until itemsArr becomes not empty.
Is your get-data module doing some browser things with protractor? If so, you will need to set/get itemsArr within the context of the controlFlow. Otherwise it will read all the code in the get-data module, but defer its execution and not wait for it to finish before moving right along to those expect statements.
var data = require('get-data'); //custom module here
var itemsArr;
describe('Test', function() {
beforeAll(function() {
// hook into the controlFlow and set the value of the variable
browser.controlFlow().execute(function() {
data.get(function(err, result) {
itemsArr = result; //load data from module
});
});
});
//error: Cannot read property 'forEach' of undefined
describe('check each item', function() {
// hook into the controlFlow and get the value of the variable (at that point in time)
browser.controlFlow().execute(function() {
itemsArr.forEach(function(item) {
checkItem(item);
});
});
});
function checkItem (item) {
var itemName = item.name;
describe(itemName, function() {
console.log('describe');
it('should work', function() {
console.log('it');
expect(true).toBeTruthy();
});
});
}
});
I'm writing a few tests for an Angular application, these are my first stab at unit tests for Angular using Jasmine. I'm having trouble structuring the test to cater for the various scenarios inside the function (namely the if statement and callbacks).
Here's my $scope function, which takes an Object as an argument, and if that object has an id, then it updates the object (as it'll already exist), otherwise it creates a new report and pushes to the backend using the CRUD service.
$scope.saveReport = function (report) {
if (report.id) {
CRUD.update(report, function (data) {
Notify.success($scope, 'Report updated!');
});
} else {
CRUD.create(report, function (data) {
$scope.report = data;
Notify.success($scope, 'Report successfully created!');
});
}
};
My test so far passes in a fake Object with an id so it'll trigger the CRUD.update method, which I then check is called.
describe('$scope.saveReport', function () {
var reports, testReport;
beforeEach(function () {
testReport = {
"id": "123456789",
"name": "test"
};
spyOn(CRUD, 'update');
$scope.saveReport(testReport);
});
it('should call CRUD factory and update', function () {
expect(CRUD.update).toHaveBeenCalledWith(testReport, jasmine.any(Function));
});
});
I understand Jasmine doesn't allow multiple spies, but I want to be able to somehow test for the if condition, and run a mock test for when the Object doesn't pass in an Object too:
describe('$scope.saveReport', function () {
var reports, testReport;
beforeEach(function () {
testReport = {
"id": "123456789",
"name": "test"
};
testReportNoId = {
"name": "test"
};
spyOn(CRUD, 'update');
spyOn(CRUD, 'create'); // TEST FOR CREATE (NoId)
spyOn(Notify, 'success');
$scope.saveReport(testReport);
$scope.saveReport(testReportNoId); // TEST FOR NO ID
});
it('should call CRUD factory and update', function () {
expect(CRUD.update).toHaveBeenCalledWith(testReport, jasmine.any(Function));
// UNSURE ON THIS PART TOO
});
});
I've read things about using the .andCallFake() method, but I could not see how this could work with my setup. Any help really appreciated.
It seems that you should decide on what you need to test first. If you want to test simply that update is called when id exists or create is called when it does not then you should just structure the it function with those conditions. The before each is the wrong place for some of those things.
it('should call CRUD factory and update', function () {
spyOn(CRUD, 'update');
$scope.saveReport(testReport);
expect(CRUD.update).toHaveBeenCalledWith(testReport, jasmine.any(Function));
});
it('should call CRUD create', function() {
spyOn(CRUD, 'create');
$scope.saveReport(testReportNoId); // TEST FOR NO ID
expect(CRUD.create).toHaveBeenCalledWith(testReport, jasmine.any(Function));
});
Only put things in the before each that you actually should do before each test.
Hope this helped!
I would like to run Protractor tests on two separate pages in my Angular app: /dashboard and /articles.
The complication is that I have to log in to the app manually.
Currently I have this setup:
var LoginPage = function() {
ptor = protractor.getInstance();
this.login = function(url) {
ptor.get(url);
ptor.findElement(protractor.By.model('email')).sendKeys(config.LOGIN_EMAIL);
ptor.findElement(protractor.By.model('password')).sendKeys(config.LOGIN_PASS);
ptor.findElement(protractor.By.tagName('button')).click();
};
};
describe('The dashboard', function() {
console.log('logging in');
var loginPage = new LoginPage();
loginPage.login(config.DASHBOARD_URL);
console.log('logged in');
it('has a heading', function() {
console.log('testing dashboard 1');
heading = ptor.findElement(protractor.By.tagName('h1'));
expect(heading.getText()).toEqual(config.DASHBOARD_HEADING);
});
});
describe('The article widget', function() {
console.log('logging in');
var loginPage = new LoginPage();
loginPage.login(config.ARTICLE_URL);
console.log('logged in');
it('has a heading', function() {
console.log('testing article 1');
heading = ptor.findElement(protractor.By.tagName('h1'));
expect(heading.getText()).toEqual(config.ARTICLES_HEADING);
});
});
This gives me the following output:
Selenium standalone server started at http://192.168.2.9:56791/wd/hub
logging in
LoginPage
logged in
logging in
LoginPage
logged in
testing dashboard 1
Ftesting article 1
It looks as though both the describe sections are kicking off in parallel. How can I force the following sequence of events, while still structuring the code in a sensible way?
Load dashboard page
Log in
Run dashboard tests
Load article page (Assume we are already logged in)
Run article tests
You can move the login to another file.
Then, in your protractor configuration file do this:
exports.config = {
specs: [
'spec/login.js',
'spec/dashboard_test.js',
'spec/article_test.js'
],
...
};
Login will run before the other tests
describe('my app', function(){
beforeEach(function(){
login()...
})
describe('dashboard');
describe('the article widget')
});
The Protractor documentation recommends
put your log-in code into an onPrepare function, which will be run once before any of your tests.
For example in your protractor.conf
onPrepare: function() {
browser.driver.get('http://localhost/login.html');
browser.driver.findElement(by.id('username')).sendKeys('Jane');
browser.driver.findElement(by.id('password')).sendKeys('1234');
browser.driver.findElement(by.id('clickme')).click();
// Login takes some time, so wait until it's done.
// For the test app's login, we know it's done when it redirects to
// index.html.
return browser.driver.wait(function() {
return browser.driver.getCurrentUrl().then(function(url) {
return /index/.test(url);
});
}, 10000);
}
I had a similar issue with my e2e protractor tests. Describe blocks were being executed in parallel, causing my tests to fail.
My code before the fix was something like:
describe('test1', function() {
it('do foo1', function() {..});
describe('do test1', function() {..});
});
describe('test2', function() {
it('do foo2', function() {..});
describe('do test2', function() {..});
});
Both the describe blocks were being executed in parallel causing my tests to fail. The fix was to enclose the it blocks in describe blocks.
Code after the fix:
describe('test1', function() {
describe('foo1', function() {
it('do foo1', function() {..});
});
describe('do test1', function() {..});
});
describe('test2', function() {
describe('foo2', function() {
it('do foo2', function() {..});
});
describe('do test2', function() {..});
});
Link to a similar issue on protractor github: https://github.com/angular/protractor/issues/592
I am trying to test a function that requires a module using jasmine and requirejs.
Here is a dummy code:
define("testModule", function() {
return 123;
});
var test = function() {
require(['testModule'], function(testModule) {
return testModule + 1;
});
}
describe("Async requirejs test", function() {
it("should works", function() {
expect(test()).toBe(124);
});
});
It fails, because it is an async method. How can I perform a test with it?
Note: I dont want to change my code, just my tests describe function.
For testing of an asynchronous stuff check runs(), waits() and waitsFor():
https://github.com/pivotal/jasmine/wiki/Asynchronous-specs
Though this way looks a bit ugly as for me, therefore you could also consider following options.
1. I'd recommend you to try jasmine.async that allows you to write asynchronous test cases in this way:
// Run an async expectation
async.it("did stuff", function(done) {
// Simulate async code:
setTimeout(function() {
expect(1).toBe(1);
// All async stuff done, and spec asserted
done();
});
});
2. Also you can run your tests inside require's callback:
require([
'testModule',
'anotherTestModule'
], function(testModule, anotherTestModule) {
describe('My Great Modules', function() {
describe('testModule', function() {
it('should be defined', function() {
expect(testModule).toBeDefined();
});
});
describe('anotherTestModule', function() {
it('should be defined', function() {
expect(anoterTestModule).toBeDefined();
});
});
});
});
3. Another point is I guess that this code is not working as you're expecting:
var test = function() {
require(['testModule'], function(testModule) {
return testModule + 1;
});
};
Because if you call test(), it won't return you testModule + 1.