PhantomJS Runners for Mocha

Mocha is a feature-rich JavaScript test framework running on node and the browser. Along with the Chai assertion library they make an impressive combo. PhantomJS is a headless WebKit with a JavaScript/CoffeeScript API. It has fast and native support for various web standards like DOM handling, CSS selectors, JSON, Canvas, and SVG.

The mocha-phantomjs project provides a mocha-phantomjs.coffee script file and extensions to drive PhantomJS while testing your HTML pages with Mocha from the console. The preferred usage is to install mocha-phantomjs via node's packaged modules and use the mocha-phantomjs binary wrapper. Tested with Mocha 1.12.x, Chai 1.7.x, and PhantomJS 1.9.1.

• Since version 3.0 of mocha-phantomjs, you must use PhantomJS 1.9.1 or higher.

Key Features

Standard Out

Finally, process.stdout.write, done right. Mocha is primarily written for node, hence it relies on writing to standard out without trailing newline characters. This behavior is critical for reporters like the dot reporter. We make up for PhantomJS's lack of stream support by both customizing console.log and creating a process.stdout.write function to the current PhantomJS process. This technique combined with a handful of fancy ANSI cursor movement codes allows PhantomJS to support Mocha's diverse reporter options.

Exit Codes

Proper exit status codes from PhantomJS using Mocha's failures count. So in standard UNIX fashion, a 0 code means success. This means you can use mocha-phantomjs on your CI server of choice.

Mixed Mode Runs

You can use your existing Mocha HTML file reporters side by side with mocha-phantomjs. This gives you the option to run your tests both in a browser or with PhantomJS. Since mocha-phantomjs needs to control when the run() command is sent to the mocha object, we accomplish this by setting the mochaPhantomJS on the window object to true. Below, in the usage section, is an example of a HTML structure that can be used both by opening the file in your browser or choice or using mocha-phantomjs.

Installation

We distribute mocha-phantomjs as an npm that is easy to install. Once done, you will have a mocha-phantomjs binary. See the next usage section for docs or use the -h flag.

Since 3.4, we now declare phantomjs as a peer dependency, and it will be installed adjacent to mocha-phantomjs automatically. You may use -p to provide an explicit path to phantomjs, or call phantomjs directly yourself via phantomjs lib/mocha-phantomjs.coffee <page> <reporter> <config-as-JSON>. The later approach is recommended for build system plugins to avoid another process fork.

Due to a bug in phantomjs 1.9.8+, mocha-phantomjs currnently caps its peer depedency to 1.9.7-15. If you are having install errors due to peer depdendency resolution conflics, expliclity declare your dependency as "phantomjs": "1.9.7-15", or expliclity install 1.9.7, like npm install phantomjs@1.9.7-15.

Usage

Usage: mocha-phantomjs [options] page

 Options:

   -h, --help                   output usage information
   -V, --version                output the version number
   -R, --reporter <name>        specify the reporter to use
   -f, --file <filename>        specify the file to dump reporter output
   -t, --timeout <timeout>      specify the test startup timeout to use
   -g, --grep <pattern>         only run tests matching <pattern>
   -i, --invert                 inverts --grep matches
   -b, --bail                   exit on the first test failure
   -A, --agent <userAgent>      specify the user agent to use
   -c, --cookies <Object>       phantomjs cookie object http://git.io/RmPxgA
   -h, --header <name>=<value>  specify custom header
   -k, --hooks <path>           path to hooks module
   -s, --setting <key>=<value>  specify specific phantom settings
   -v, --view <width>x<height>  specify phantom viewport size
   -C, --no-color               disable color escape codes
   -p, --path <path>            path to PhantomJS binary


 Examples:

   $ mocha-phantomjs -R dot /test/file.html
   $ mocha-phantomjs http://testserver.com/file.html
   $ mocha-phantomjs -s localToRemoteUrlAccessEnabled=true -s webSecurityEnabled=false http://testserver.com/file.html
   $ mocha-phantomjs -p ~/bin/phantomjs /test/file.html

Now as an node package, using mocha-phantomjs has never been easier. The page argument can be either a local or fully qualified path or a http or file URL. --reporter may be a built-in reporter or a path to your own reporter (see below). See phantomjs WebPage settings for options that may be supplied to the --setting argument.

Your HTML file's structure should look something like this. The reporter set below to html is only needed for viewing the HTML page in your browser. The mocha-phantomjs.coffee script overrides that reporter value. The conditional run at the bottom allows the mixed mode feature described above.

<html>
  <head>
    <meta charset="utf-8">
    <!-- encoding must be set for mocha's special characters to render properly -->
    <link rel="stylesheet" href="mocha.css" />
  </head>
  <body>
    <div id="mocha"></div>
    <script src="mocha.js"></script>
    <script src="chai.js"></script>
    <script>
      mocha.ui('bdd');
      mocha.reporter('html');
      expect = chai.expect;
    </script>
    <script src="test/mycode.js"></script>
    <script>
      if (window.mochaPhantomJS) { mochaPhantomJS.run(); }
      else { mocha.run(); }
    </script>
  </body>
</html>

Screenshots

Mocha-phantomjs supports creating screenshots from your test code. For example, you could write a function like below into your test code.

function takeScreenshot() {
  if (window.callPhantom) {
    var date = new Date()
    var filename = "screenshots/" + date.getTime()
    console.log("Taking screenshot " + filename)
    callPhantom({'screenshot': filename})
  }
}

If you want to generate a screenshot for each test failure you could add the following into your test code.

  afterEach(function () {
    if (this.currentTest.state == 'failed') {
      takeScreenshot()
    }
  })

Supported Reporters

Mocha-phantomjs does not scrap the web page under test! No other PhantomJS driver stacks up to our runner support. Some have used a debounce method to keep duplicate messages in the spec reporter from showing up twice. Loosing one of Mocha's console reporters neatest features, initial test start feedback. The animation below is an example of how our runner script fully compiles with expected Mocha behavior.

The following is a list of tested reporters. Since moving PhantomJS 1.9.1, most core Mocha reporters should "just work" since we now support stdout properly. Reporters with node dependencies will not work, like html-cov. If you have an issue with a reporter, open a github issue and let me know.

Spec Reporter (default)

The default reporter. You can force it using spec for the reporter argument.

Dot Matrix Reporter

Use dot for the reporter argument.

The PhantomJS process has no way of knowing anything about your console window's width. So we default the width to 75 columns. But if you pass down the COLUMNS environment variable, it will pick that up and adjust to your current terminal width. For example, using the $COLUMNS variable like so.

env COLUMNS=$COLUMNS phantomjs mocha-phantomjs.coffee URL dot

Bundled and tested reporters include:

tap
min
list
doc
teamcity
json
json-cov
xunit
progress
landing
markdown

Third Party Reporters

Mocha has support for custom 3rd party reporters, and mocha-phantomjs does support 3rd party reporters, but keep in mind - the reporter does not run in Node.js, but in the browser, and node modules can't be required. You need to only use basic, vanilla JavaScript when using third party reporters. However, some things are available:

• require: You can only require other reporters, like require('./base') to build off of the BaseReporter
• exports, module: Export your reporter class as normal
• process: use process.stdout.write preferrably to support the --file option over console.log (see #114)

Also, no compilers are supported currently, so please provide JavaScript only for your reporters.

Testing

Simple! Just clone the repo, then run npm install and the various node development dependencies will install to the node_modules directory of the project. If you have not done so, it is typically a good idea to add /node_modules/.bin to your $PATH so these modules bins are used. Now run npm test to start off the test suite.

We also use Travis CI to run our tests too. The current build status:

Alternatives

• OpenPhantomScripts - https://github.com/mark-rushakoff/OpenPhantomScripts
• Front Tests - https://github.com/Backbonist/front-tests

License

Released under the MIT license. Copyright (c) 2012 Ken Collins.