NOTE: This project is forked from some work done by Ivan DeMarino here. He also wrote an article describing his approach. The only change I made was to add the same validation.js and test file that I'm using in the other sample projects.

-RC

This is an example project layout, that realises JavaScript Unit Testing using Jasmine on the top of PhantomJS.

It is designed to be used standalone or within Maven.

$ phantomjs test/phantomjs_jasminexml_runner.js test/test_runner.html xml_output_dir/

This will produce a set of XML outputs, one for every Jasmine Test Suite.

The project already provides a pom.xml that does the trick. So a simple:

$ mvn test

or

$ mvn clean install

will launch the tests and, if they all pass, make the mvn build pass. If any test fails, the mvn build fails.

Well, explaining the internals of PhantomJS here is pointless, but to understand how this works, you need to understand a simple idea: PhantomJS provides a main "Javascript Context" from within which, the user is able to spawn a WebPage, effectively creating another "Javascript Context".

The new context has the following characteristics:

• isolated - the page doesn't know anything about phantom and phantom API
• controllable - the main phantom context provides API to control/influence the page context

The magic is done by 2 scripts:

test/phantomjs_jasminexml_runner.js
test/lib/jasmine-reporters/jasmine.phantomjs-reporter.js

jasmine.phantomjs-reporter.js is a plugin for Jasmine that saves the result of the tests at the bottom of the webpage where the test runs.

phantomjs_jasminexml_runner.js instead runs the show, launches the tests, extracts the result from the page, then saves it to the directory passed (see source).

Happy Testing ;-)