NOTE: this post was written for Karma 0.8 which required a manual installation of PhantomJS.
However this blog post is still relevant for installing the NodeJS and NPM pre-requisites.
As of 0.10 both PhantomJS and Chrome will be automatically installed by the launcher plugins.
Installation instructions for Karma 0.10 can be found here (a "Local installation" is preferred).
Furthermore instructions on how to install plugins (introduced as of 0.10) can be found here.

Recently I decided to switch from the “Jasmine Maven Plugin” (using the Mozilla Rhino JavaScript “emulator”) to the Karma (previously called Testacular) test runner. The big advantage of Karma opposed to the “Jasmine Maven Plugin” is that it uses actual browsers (like Chrome, Firefox, Safari and even IE) to execute the tests. This blogpost describes the installation and configuration of Karma on Windows. To install Karma you first need to install NodeJS and its NPM (NodeJS Package Manager). Additionally you could install PhantomJS, a “headless” web-kit browser, to run your JavaScript tests from the command-line without spawning unwanted browser windows. Instead of using PhantomJS as a replacement for testing against real browsers you could use it to run tests on your local development machine while your continuous integration server could then run your tests on “all” relevant browsers.

Continue reading →

When experimenting with AngularJS features I often create a so-called Plunk on http://plnkr.co/. Although this site provides built-in templates for AngularJS, I found it useful to create my own since:

• The “1.0.x (stable)” / “1.0.2 + Jasmine” templates use old versions of AngularJS;
• “1.0.2 + Jasmine” template solely outputs Jasmine results but no AngularJS content.

Continue reading →

Next to creating controllers and directives, AngularJS also supports “singleton” services. Services, like on the server-side, offer a great way for separating logic from your controllers. In AngularJS anything that’s either a primitive type, function or object can be a service. Although the concept of service is quite straight forward, the declaration of them in AngularJS isn’t:

• There are 4 different ways to declare a service.
  • Registering a existing value as a service
  • Registering a factory function to create the singleton service instance
  • Registering a constructor function to create the singleton service instance
  • Registering a service factory which can be configured
• Only 1 of them is extensively documented

Continue reading →

Having a history as a Java developer I prefer declaring the complete JavaScript object at once through an object literal; in a similar fashion as your would declare a class in Java. In my opinion adding new properties "on the fly" to a JavaScript object is a very bad practice:

var jsLibrary = { name: 'AngularJS' };
// adds a new property "homepage" to the existing object...
jsLibrary.homepage = 'http://www.angularjs.org/';

Continue reading →

AngularJS is an excellent JavaScript web framework offering so-called "directives" to 'teach' HTML some new tricks. Examples of built-in AngularJS directives are:

• "ngView": defines the placeholder for rending views
• "ngModel": binds scope properties to "input", "select" and "text" elements
• "ngShow" / "ngDisabled": for showing or disabling an element based on the result of an expressions

Continue reading →

Like most Java developers I used to have a serious aversion to JavaScript. I was quite happy to delegate any 'scripting' stuff to fellow developers. At my current project, we initially decided to use the Vaadin web framework. It seemed the perfect choice for creating Rich Internet Application (RIA) user-interfaces without writing a single line of JavaScript. However what originally seemed to be a sensible choice, turned out to be a dead-end:

• Vaadin is highly dependent on http sessions and as it turns out doesn't play well when being clustered.
• No default support for server push; also the 'most stable' Vaadin add-on turned out to be quite unstable and incompatible with clustering.
• No wrapper existed for v3 of Google Maps; as an alternative we used the OpenLayers Add-on instead. However this add-on turned out to be not so stable either and lacked the user experience of Google Maps to which users are accustomed to (like dragging the 'pegman' on the map in order to show street view).

Continue reading →

Recently I've been looking into Spring Roo to find ways to speed-up software development as well as reducing plumbing code. In this blog post I will highlight some of the advantages and disadvantages of Spring Roo which I stumbled upon. So don't expect a full featured introduction or tutorial; plenty of those already exist on the internet ;-)

Spring Roo takes on a different approach to code generation compared to other solutions. Instead of generating additional .java files (through the "Generation Gap Pattern") it generates so-called AspectJ inter-type declaration (ITD) .aj source files. Each generated inter-type declaration (ITD) type will "weave in" structural changes to its target .java file; for example to add new methods (i.e. getter/setter methods) or an "implements ..." clause. The big advantages of such inter-type declaration (ITD) are:

Continue reading →