Here is the documentation for Function.js.
.create automatically wraps the function into another one with the defined set of functionalities, therefore shortens your code. It's used throughout mootools itself, as the foundation for .pass, .bind, .delay, .periodical.
You probably won't find find yourself using create directly that often (well, at least I don't).
.create takes a handful of options in a single argument as an object:
By way of example, here's what Function.pass (the next item in the tutorial) looks like:
pass: function(args, bind){ return this.create({'arguments': args, 'bind': bind}); }
More on .pass in a second, but what this demonstrates is taking a function, setting up some definitions (arguments, bind) and returning it ready to execute later.
/*syntax: Function.pass(arguments, bind) bind is optional; more on bind later example: myFunction.pass([arg1, arg2], myElement);*/ function say(msg){ alert(msg); }; var insSay=say.pass('hello world'); //Do something insSay();
You might ask yourself why do this, but you'll see why in examples later. Basically, you don't always want to execute a function immediately, and without something like the .pass method, you have no way to pass variables to that function without executing it. myFunction(value) executes myFunction. myFunction.pass(value) sets up myFunction so that it can execute later.
Let's say you have an effect that's going to make something fade out. When it's done, you want to have it execute a callback of some sort. Your effect has an option for you to hand it a function to execute when it's finished, but you need to pass that function some variables. If you did something like: "onComplete: myFunction(value)" you'll execute your function immediately, and onComplete will be equal to whatever is returned, which isn't what you want. Instead, you do "onComplete: myFunction.pass(value)" - which gets you want you want. myFunction.pass(value) returns an instance of that function with that parameter set. So when you execute onComplete() you're really executing myFunction(value).
The "this" namespace inside an object (such as a function) refers to the instance of that object. It's really useful for a lot of things, not the least of which is referring to other functions in the same object or the state of some variable. The problem is that you might author functionality that needs to be able to refer to variables in the object that's calling it. Below is a simple example of this, but when you write more complex classes and functions you'll find that this ability to define what "this" is in a given context is a necessity.
This example should change this paragraph to have a border around it.
function myFunction(){ this.setStyle('border', '1px solid black'); /*note that 'this' here refers to myFunction, not an element we'll need to bind this function to the element we want to alter*/ }; /*bind the element I want to effect to the function note that a function is returned, not executed.*/ var myBoundFunction = myFunction.bind($('bindExample')); /*execute that new function with the bound object*/ myBoundFunction();
You use bindAsEventListener when you want as the first argument the event, no matter if its ie or firefox or any other browser.
function fn(event){ //event is the event, no need to do window.event || event //the keyword this is the element, because we used element in the bindAsEventListener argument. }; element.onclick = fn.bindAsEventListener(element);
with bindWithEvent you'll get the mootools Event wrapper as the first argument:
function fn(event){ //event is the mootools event, that supports all the cool stuff like event.page.y //position of the mouse in the page event.key //the key pressed event.shift //if shift has been pressed to trigger the event //and tons of others, refer to the docs, Native/Event.js //the keyword this is the element, because we used element in the bindWithEvent argument. }; element.onclick = fn.bindWithEvent(element);
Basically bind withEvent will do this automatically for you:
function fn(event){ event = new Event(event); }; element.onclick = fn;
Yes, you can set up delays with setTimeout, but it's sometimes convoluted and complicated. Additionally, delay allows you to bind an element to the function that's going to get delayed.
The example below should draw a border around this paragraph after a one second delay.
function myFunction(){ this.setStyle('border', '1px solid black'); }; myFunction.delay(1000, $('delayExample'));
Here's a great example of binding. If the element were an argument to the function, I couldn't call the function with an argument with out executing it. This type of syntax wouldn't work:
myFunction($('delayExample')).delay(500);
...couldn't work, because you'd execute the function immediately, and .delay would then attempt to execute as a method on what was returned.
You could do this though:
myFunction.delay(500, $('delayExample')); //OR myFunction.bind($('delayExample')).delay(500);
.delay takes three parameters: the delay (in ms), an object to bind (can be null), and any arguments you wish to pass the function. If you only pass one, you can just specify the argument, if you pass more than one, use an array:
myFunction.delay(500, null, [$('delayExample'), argument2, etc.]);
There's also //.periodical//, which repeatedly executes your function.
function blink(element){ if(element.getStyle('visibility') == 'visible') element.setStyle('visibility', 'hidden'); else element.setStyle('visibility', 'visible'); } var blinker = blink.periodical(200, null, $('periodicalExample')); /*curses! the blink tag is back!*/ $clear.delay(2200, null, blinker);//make it stop!
This is a shortcut for try/catch. Function.attempt will catch any errors in your function and return false to you if the function fails (otherwise it returns the results of your function).
This is a change from MooTools 1.0 where Function.attempt() returned the error thrown instead.
function test(value){ alert(value.something); /*in this test, value won't be defined, so it'll throw an error*/ } test.attempt(); /*returns false because of the error*/
The above is the equivalent of this:
function test(value){ try { alert(value.something); /*in this test, value won't be defined, so it'll throw an error*/ } catch(err) { return false; } }
JavaScript has a native method to run a function and bind something to it (fun.apply(bind, arrayOfArguments)). It also has a native method that does the same thing with the arguments specified after the bind object (fun.call(bind[, arg1[, arg2[, ...]]]);). You can always just call a function and pass arguments (foo(bar)).
It's basically similar to call or apply as it executes the function immediately but the it will be bound to the second argument. If the first argument is an array then the original function will get its elements as arguments, if it is not an array then the function will only get that single argument. It's like a reverse apply that accepts anything not only an array.
foo.run('hi there'); //same as foo('hi there'); foo.run('hi there', this); //same as foo.apply(this, ['hi there']); //also same as foo.call(this, 'hi there'); foo.run(['hi', 'there'], this); //same as foo.apply(this, ['hi', 'there']); //also same as foo.call(this, 'hi', 'there');