Organizing Test Suites

While your project is growing you will inevitably add more and more integration tests. This will increase your build time and will also slow down your productivity. To prevent this you should start to run your tests in parallel. You might have already recognised that WebdriverIO creates for each spec (or feature file in cucumber) a single Selenium session. In general, you should try to test a single feature in your app in one spec file. Try to not have too many or too less tests in one file. However, there is no golden rule about that.

Once you get more and more spec files you should start running them concurrently. To do so you can adjust the maxInstances property in your config file. WebdriverIO allows you to run your tests with maximum concurrency meaning that no matter how many files and tests you have, they could run all in parallel. Though there are certain limits (computer CPU, concurrency restrictions).

Let’s say you have 3 different capabilities (Chrome, Firefox, and Safari) and you have set maxInstances to 1, the wdio test runner will spawn 3 processes. Therefore, if you have 10 spec files and you set maxInstances to 10; all spec files will get tested at the same time and 30 processes will get spawned.

You can define the maxInstance property globally to set the attribute for all browser. If you run your own Selenium grid it could be that you have more capacity for one browser than for an other one. In this case you can limit the maxInstance in your capability object:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
// wdio.conf.js
exports.config = {
// ...
// set maxInstance for all browser
maxInstances: 10,
// ...
capabilities: [{
browserName: "firefox"
}, {
// maxInstances can get overwritten per capability. So if you have an in-house Selenium
// grid with only 5 firefox instance available you can make sure that not more than
// 5 instance gets started at a time.
browserName: 'chrome'
}],
// ...
}

Inherit From Main Config File

If you run your test suite in multiple environments (e.g. dev and integration) it could be helpful to have multiple configuration files to keep them easy manageable. Similar to the page object concept you first create a main config file. It contains all configurations you share across environments. Then for each environment you can create a file and supplement the information from the main config file with environment specific ones:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
// wdio.dev.config.js
var merge = require('deepmerge');
var wdioConf = require('./wdio.conf.js');

// have main config file as default but overwrite environment specific information
exports.config = merge(wdioConf.config, {
capabilities: [
// more caps defined here
// ...
],

// run tests on sauce instead locally
user: process.env.SAUCE_USERNAME,
key: process.env.SAUCE_ACCESS_KEY,
services: ['sauce']
});

// add an additional reporter
exports.config.reporters.push('allure');

Group Test Specs

You can easily group test specs in suites and run single specific suites instead of all of them. To do so you first need to define your suites in your wdio config:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
// wdio.conf.js
exports.config = {
// define all tests
specs: ['./test/specs/**/*.spec.js'],
// ...
// define specific suites
suites: {
login: [
'./test/specs/login.success.spec.js',
'./test/specs/login.failure.spec.js'
],
otherFeature: [
// ...
]
},
// ...
}

If you now want to run a single suite only you can pass the suite name as cli argument like

1
$ wdio wdio.conf.js --suite login

or run multiple suites at once

1
$ wdio wdio.conf.js --suite login,otherFeature

Run Single Test Suites

If you are working on your WebdriverIO tests you don’t want to execute your whole suite everytime you added an assertion or any other code. With the --spec parameter you can specify which suite (Mocha, Jasmine) or feature (Cucumber) should be run. For example if you only want to run your login test, do:

1
$ wdio wdio.conf.js --spec ./test/specs/e2e/login.js

Note that each test file is running in a single test runner process. Since we don’t scan files in advance you can’t use for example describe.only at the top of your spec file to say Mocha to only run that suite. This feature will help you though to do that in the same way.

Stop testing after failure

With the bail option you can specify when WebdriverIO should stop the test run after test failures. This can be helpful when you have a big test suite and want to avoid long test runs when you already know that your build will break. The option expects a number that specifies after how many spec failures it should stop the whole test run. The default is 0 meaning that it always runs all tests specs it can find.