Unit testing in WaSaby environment

Introduction

It's based on Mocha and Chai. Sinon is also included.

There are some global functions which available in test cases: describe, it and many more from Mocha's BDD interface.

You can include assert method like in this example.

All test cases should be named by mask *.test.js. For example, test/example.test.js:

   /* global describe, context, it */
   import {assert} from './assert';
   import {MyModule} from '../MyPackage/MyLibrary';

   describe('MyPackage/MyLibrary:MyModule', () => {
      let myInstance;

      beforeEach(() => {
         myInstance = new MyModule();
      });

      afterEach(() => {
         myInstance = undefined;
      });

      describe('.constructor()', () => {
         it('should return instance of MyModule', () => {
            assert.instanceOf(myInstance, MyModule);
         });
      });
   });

The first thing

Add saby-units as development dependency in package.json:

  "devDependencies": {
    "saby-units": "git+https://git.sbis.ru:saby/Units.git#rc-3.19.100"
  }

And install it:

npm install

All files in examples below should be created in the root directory of your package.

Configure

Add build script in scripts section of package.json:

  "scripts": {
    "build": "saby-units --install"
  }

This script creates links defined in saby-units section below.

Add saby-units section to your package.json file, for example:

  "saby-units": {
    "root": "./src"
  }

The default config looks like:

  "saby-units": {
    "moduleType": "esm",
    "root": ".",
    "tests": "",
    "report": "artifacts/xunit-report.xml",
    "htmlCoverageReport": "/artifacts/coverage/index.html",
    "jsonCoverageReport": "artifacts/coverage.json",
    "timeout": 10000,
    "links": {},
    "dependencies": [],
    "patchedRequire": false,
    "url": {
      "scheme": "http",
      "host": "localhost",
      "port": 80,
      "path": "",
      "query": "reporter=XUnit"
    }
  }

Parameters explanation:

• moduleType: type of module to test, 'esm' or 'amd';
• root: document root for testing files;
• tests: folder where tests are (if placed in separated folder);
• report: location of report in XUnit format which will created during testing;
• htmlCoverageReport: location of report in HTML format that created by nyc package;
• jsonCoverageReport: location of report in JSON format that created by nyc package;
• timeout: timeout to waiting for testing ends;
• links: create symbolic links to folders, a map looks like 'source folder' -> 'link'. Example:

"links": {
  "node_modules/saby-types/Types": "Types"
}

• dependencies: AMD modules to load beore testing starts. Example:

"dependencies": ["testing/init"]

• patchedRequire: Use patched version of RequireJS;
• url: parts of URL of testing app to locate in browser.

You don't need to copy all the parameters to your own, you should set only changed. Preferred server port, for example.

Run under Node.js

1. Add script to scripts section in your package.json file:

   "test": "saby-units --isolated"

2. And run it:

    npm run test
   

You can save the report in XML format if you want to:

"test:report": "saby-units --isolated --report"

Debugging in Visual Studio Code

1. Create a new launch configuration with this config.

Create coverage report under Node.js

1. Add to package.json section with setting for nyc package, here are recommended settings below (you just need to replace patterns in include section):

     "nyc": {
       "include": [
         "Foo/**/*.js",
         "Bar/**/*.js"
       ],
       "reporter": [
         "text",
         "html"
       ],
       "cache": false,
       "eager": true,
       "report-dir": "./artifacts/coverage"
     }

2. Add script to scripts section in your package.json file:

   "test:coverage": "saby-units --isolated --coverage"

3. And run it:

    npm run test:coverage
   

There are some important keys for nyc:

• include: mask for files to include in coverage;
• reporter: format of the coverage report;
• report-dir: path to folder to put the report to.

You can find out more information about fine tune at nyc's site.

An important notice about symbolic links in instrumented code

If you have symbolic links in folder included for instrumentation you must mind the unexpected nyc behaviour which prevent these files and folders from being instrumented. As a result you don't see desired files in coverage report. To solve this problem you should use undeclared cwd key which sets the default working dir for nyc:

   "nyc": {
     "cwd": "./",

Also you probably have to revise report-dir value according to this change.

Run via chrome

Headless Chrome

1. Add script to scripts section in your package.json file:

     {
        "test:browser": "saby-units --browser"
     }

2. And run it:

    npm run test:browser
   

Headed Chrome

1. Add script to scripts section in your package.json file:

     {
        "test:chrome": "saby-units --browser --head"
     }

2. And run it:

    npm run test:chrome
   

Run via Selenium webdriver

1. Add script to scripts section in your package.json file:

     {
        "test:selenium": "saby-units --browser --selenium"
     }

2. And run it:

    npm run test:selenium
   

Run testing server only

You can only run the testing server and check tests in browser manually.

1. Add script to scripts section in your package.json file:

     {
        "test:app": "saby-units --server"
     }

2. And run it:

    npm run test:app
   

3. Open your web brower and navigate to testing page (you should change the port number if you've changed it in package.json file).

Integation with Jenkins

There are some setting you have to define. In samples below we will use out npm scripts from above.

'Source code' section

✓ Multiple SCMs

+GIT:

    Repository URL: [email protected]:your/module.git

    Credentials: gitread

    Branches to build: */master

    Additional Behaviours:

        +Advanced clone behaviours

            ✓ Shallow clone

'Environment' section

✓ Inject environment variables to the build process

There are available environment variables:

WEBDRIVER_remote_enabled - run on remote Selenium grid (0 by default; change to 1 if you want to use remote selenium grid. Also you have to change host name in URL at testing-browser.js instead of localhost)

WEBDRIVER_remote_hostname - host name where Selenium grid is available (localhost by default)

WEBDRIVER_remote_port - port where Selenium grid is availbale (4444 by default)

WEBDRIVER_remote_capabilities_browserName - browser name to run test cases in (chrome by default)

WEBDRIVER_remote_capabilities_version - browser version to run test cases in

✓ Abort the build if it's stuck

Timeout minutes: 10
Time-out actions: Abort the build

'Build' section

+Run shell script (to run testing under Node.js and to generate the coverage report):

#npm config set registry http://npmregistry.sbis.ru:81/
npm install
npm run test:coverage
npm run test:report

+Run shell script (to run testing via webdriver)

npm install
npm run test:browser

'After build operations' section

Publish JUnit test result report

Add path to the XML report: artifacts/xunit-report.xml

✓ Retain long standard output/error

The path depend on settings you set in files above.

Publish documents

Title: Coverage report

Directory to archive: artifacts/coverage/lcov-report/

The path depend on settings you set in package.json.