2010 New Year’s Resolution
I have a backlog of draft posts, and I’m going to start working my way through them. I had a very interesting semester, and the next should be even more interesting. I’m keen to resume sharing. Look for more posts to come.
I have a backlog of draft posts, and I’m going to start working my way through them. I had a very interesting semester, and the next should be even more interesting. I’m keen to resume sharing. Look for more posts to come.
Our JavaScript Developer Survey results showed that a large percentage of readers don’t currently unit test their JavaScript. I’ve been unit testing my own JavaScript projects for a few years, which resulted in the creation of riotjs — for more on riotjs see my blog post: Riotjs: A JavaScript Unit Testing Framework
There are lots of JavaScript test frameworks. I’ve presented a summary below — there are both traditional unit testing libraries and behaviour-driven frameworks.
unittest.js by Thomas Fuchs has been used by many JavaScript developers since Scriptaculous became popular in 2006. It provides a test framework that is recognisable to Ruby developers, and the relationship between Rails and Scriptaculous further increased adoption in this group.
This library depends on Prototype and is usually run inside HTML templates rather than a JavaScript interpreter.
Tests look like this:
new Test.Unit.Runner({
testExample: function() { with(this) {
var exampleVar = '1234';
assertEqual('1234', exampleVar);
assert(true);
}}
}, "testlog");
Fuchs uses with(this) to cut down unnecessary syntax — you don’t have to write tests this way though.
Setup and teardown methods are supported. Shoulda-style assertion names have also been added. This adds methods to the prototypes for common JavaScript objects, so it’s possible to write 'test'.shouldEqual('test').
unittest.js also includes a handy benchmark method, and the following assertions:
assertNotEqual(expected, actual) |
Inverse of assertEqual |
assertNotVisible(element) |
Checks if an element has display: none |
assertVisible(element) |
Inverse of above |
assertInspect(expected, actual) |
Checks if expected is the same as actual.inspect() |
assertEnumEqual(expected, actual) |
Ensured two enumerables are equal |
assertIdentical(expected, actual) |
Comparison using === |
assertNotIdentical(expected, actual) |
Inverse of above |
assertNull(obj) |
Is obj null? |
assertNotNull(obj) |
Inverse of above |
assertMatch(expected, actual) |
Performs a regex, where expected is the regex |
assertType(expected, actual) |
Checks if actual.constructor == expected |
assertNotOfType(expected, actual) |
Inverse of above |
assertInstanceOf(expected, actual) |
Checks if actual is instanceof expected |
assertNotInstanceOf(expected, actual) |
Inverse of above |
assertRespondsTo(method, obj) |
Checks if obj has a method named method |
assertReturnsTrue(method, obj) |
Ensured that the method named method called on obj returns true |
assertReturnsFalse |
Inverse of above |
assertRaise(exceptionName, method) |
Ensures calling method raises an exception named exceptionName |
assertElementsMatch([elements], expression1, expression2, ...) |
Ensures the array of elements match the list of expressions |
assertElementMatches(element, expression) |
A helper for the above assertion when supplying a single element |
JsUnitTest is a port of unittest.js without the Prototype dependency.
Development for JsUnit started in 2001, and is based on JUnit. Like unittest.js, tests are designed to be run in a web browser. By using the build.xml Ant file, tests can be run against multiple browsers on a local machine, or on multiple remote machines.
JsUnit tests are grouped within a Test Page. Test pages contain your tests. If a Test Page has setUp or tearDown functions, they will be run before and after each test function. A function called setUpPage will be run once per Test Page.
Each test function is written prefixed with test:
function testExample() {
assert(true);
}
The following assertions are available:
assert([comment], booleanValue)assertTrue([comment], booleanValue)assertFalse([comment], booleanValue)assertEquals([comment], value1, value2)assertNotEquals([comment], value1, value2)assertNull([comment], value)assertNotNull([comment], value)assertUndefined([comment], value)assertNotUndefined([comment], value)assertNaN([comment], value)assertNotNaN([comment], value)fail(comment)QUnit is used by jQuery for its unit tests. To use QUnit, include qunit.js and qunit.css within a HTML file that can display the test results.
Tests look like this:
test("a basic test example", function() {
ok(true, "this test is fine");
var value = "hello";
equals("hello", value, "We expect value to be hello");
});
The jQuery QUnit documentation has full examples.
jspec (by Yehuda Katz) is an RSpec inspired framework. Rather than traditional unit tests, jspec attempts to provide a behaviour-driven development approach.
This is best illustrated with an example:
jspec.describe("The browser environment", function() {
it("should have Array.prototype.push", function() {
Array.prototype.push.should("exist");
});
it("should have Function.prototype.apply", function() {
Function.prototype.apply.should("exist");
});
})
Tests are written with it() blocks and chained should() calls.
The GitHub repository is currently pretty spartan, with no documentation or README, but jspec is a good port for those of you who are familiar with RSpec or want a change from JUnit-style libraries.
JSpec by visionmedia (who have cropped up frequently on DailyJS) is a much more fleshed out BDD library. The author has tried to keep it minimal, but it has a unique and friendly DSL. It’s possible to write “grammar-less” tests using plain JavaScript.
Jasmine builds on the ideas of the previous libraries to provide a BDD library that supports asynchronous tests — Ajax-friendly testing.

I’ve written a few posts about canvas-based animation and graphics libraries lately. Before moving on to another topic I thought I’d mention Doodle.js, which is a sprite-based canvas library. Doodle.js simplifies working with the canvas, and includes excanvas for IE support.
Code using Doodle.js looks like this:
(function(oo) {
oo.canvas('#my_canvas');
var box = oo.rect({x:25, y:25,
width:50, height:50, fill:"#ff0000"});
box.draw(); //draw a red box with initial parameters
box.modify({fill:'rgb(0,255,0)'}).translate(50,0).draw();
box.modify({fill:'purple'}).rotate(-45).draw();
box.modify({fill:'yellow'}).translate(50,0).scale(1.5).draw();
box.modify({fill:'blue'}).transform(1.5,0,0,1.5,50,0).draw();
})($doodle);
As you can see, the API is fairly clean. There isn’t currently much documentation, but the author has written a length blog post with examples.
While Processing.js is a full drawing and animation system, and Raphael is a vector graphics library, Doodle.js provides a simple way of working with sprites.
London based Website Design Company is providing professional website design and website development services. They create a range of website applications and solutions - from secure ecommerce platforms to tailored content management system. Hampshire based Web design agency is specializing in ecommerce website development. Browse through our work portfolio on website designing. Our programming team focuses to deliver an ecommerce web design that produces great ROI. 
Our recent post about Processing.js made me think about Conway’s Game of Life. If you haven’t come across it before, it’s a simulation that works like this:
1. Any live cell with fewer than two live neighbours dies, as if caused by underpopulation.
2. Any live cell with more than three live neighbours dies, as if by overcrowding.
3. Any live cell with two or three live neighbours lives on to the next generation.
4. Any dead cell with exactly three live neighbours becomes a live cell.
I like to use this algorithm to experiment with new graphics libraries, so I had my own Processing version from back in 2005. I’ve ported it to Processing.js here conway-js.
I’ve also found lots of interesting pure JavaScript implementations. ConwayJS by Joseph Robert is my favourite. It’s pretty fast and allows you to draw cells with the mouse.
Another interesting one is game_of_life by sukhchander.
I’ve uploaded my Processing.js examples to DailyJS so you can see them without checking out the git repositories:
After yesterday’s Processing.js article I discovered this post about creating a particle system with JavaScript and a canvas: Parcycle: A Particle System with HTML5 canvas.
You can view the effect in action in this demo: Parcycle

I found the use of globalCompositeOperation very interesting in this code. It allowed the author to make particles appear lighter when they overlap by using a drawing mode rather than detecting the colour on the canvas.

Processing.js is a JavaScript/HTML canvas port of Processing. Processing provides a toolkit for programming images, animation and even games. It’s hosted in Java and provides its own mini-IDE. Processing.js can interpret Processing projects, but runs them in a browser that supports canvas.
If you’re familiar with Processing already you should be able to run some of your projects in Processing.js. Just create a simple HTML page with a canvas and include processing.js. The examples/ folder in the source distribution demonstrates how to do this using a script called init.js — this finds the canvas on the page and instantiates a processing object:
Processing(canvas, scripts[i].text);
Processing.js allows you to embed Processing scripts into HTML like this:
<script type="application/processing">
</script>

If you’re new to Processing but have a lot of JavaScript you want to reuse, you can safely call Processing from your JavaScript code. I’ve prepared an example project, fire.js which demonstrates one way of mixing JavaScript and Processing. You can pass the current Processing environment to a JavaScript class like this:
void draw() {
fire.update(this);
}
Keeping some level of separation between JavaScript code and “pure” Processing seems like a good way to avoid confusion if you ever port to another Processing platform.
Every Processing project has setup() and draw() functions. In this case I’ve instantiated a Fire object and I’m passing this which is the current Processing object. Then I can access Processing’s functions from within my JavaScript class.
The Processing reference gives example code for all of the Processing functions. There’s a lot of drawing primitives like rect() and ellipse() — playing around with these is a good way to learn Processing.
ruby-processing is a Ruby port. It differs to Processing.js because it provides a Ruby-style version of the Processing API. Even CamelCase method names have been changed to underscores.
Although Processing.js and Processing are awesome projects, if you just want to draw vector graphics you’d be better off with something like Raphael. Processing is more suited to animated visualisations or interactive projects.
This is a relatively simple concept and is nothing to elaborate but I wanted to share a small piece of code that will take every link with the class of “ajax” and access it using AJAX instead of actually going to that page. This using the same Request instance which will keep it optimized and manageable. MooTools is going to make this nice and easy on us…
var ajax_request = new Request(
{
onSuccess: function(responseText, responseXML)
{
$('message').set('html', responseText);
}
});
$$('.ajax').each(function(item){
var url = item.get('href');
item.addEvent('click', function(event){
ajax_request.options.url = url;
ajax_request.send();
return false;
});
});
Sent Via AJAX Also Sent Via AJAXI am filled with the results of our AJAX requests.
Its incredibly simple so no demo for this. The important concept here is that we are using the same request and pulling the href attribute base on a class selection. So changing something from AJAX to regular is as easy as adding or removing the class “ajax”. Pretty cool eh?
Node.js makes writing IRC-related programs straightforward and super-efficient, thanks to its evented networking. Therefore it isn’t surprising that a lot of projects have appeared already.
I’ve been looking around for reusable JavaScript IRC libraries, and discovered the following projects.
This IRC daemon demo presented at jsconf.eu/2009 demonstrates how to build a server. It has simple prototypal classes for channels and users, and parses IRC commands using a regex and a switch. It also uses node’s TCP library.
nodelog is an IRC logger. This would be a good place to start if you’d like to build a JavaScript IRC bot. The client code is clear and reusable. The log.js file demonstrates how to use the irc.js library.
irccat-nodejs is based on irccat which allows you to send messages to IRC using HTTP or a TCP socket. The client code is again fairly simple.
egret is a command-line IRC client. The base code uses node’s TCP sockets and a hash to interpret IRC commands into English.
nodejs.irc is an IRC bot. It’s not finished yet, but the code is separated out and easy to follow.
aoibot is similar to nodejs.irc, but it looks like the finished product will be less generic. Again, it’s unfinished, so I expect the author still has a lot of ideas and work to do.