How to define promises on Metro App for Windows 8 using JavaScript

Metro application from Windows 8 can be written not only in C# with XAML but also in Java Script and HTML5. As we all know we can do a lot of thinks with asynchronous programming.
In this post I want to describe how we can define asynchronous methods in Java Script. The name that is used for this kind of actions is "Promises". This can be used for operations that take a long time to complete.
A common pattern for this kind of actions is the callback pattern. Basically, when you call a method asynchron, one parameter is the callback action that is called when the action finish executing.
The definition of promises was not made my Microsoft. The CommonJS were the first one that introduces this pattern. Microsoft only adopted this pattern that is very similar to asynchronous calls of .NET library.
This pattern tries to define a common way to make asynchronous calls, in a predictable way (manner).
In a metro application, WinJS define three types of callbacks that we can use:

• complete callback
• failure callback
• progress callback

When we call an asynchronous action that support promises we can specify these three types of callbacks.

foo.someMethod().then(
 function(response){...}, // Success
 function(response){...}, // Fail
 function(status){...}) // Progress

or

foo.someMethod().then(
 successMethod, // Success
 failMethod, // Fail
 progressMethod) // Progress

Many places from WinJS library use promises, from control library to animation or binding. As you can see it is very easy to use promises, but let see what we can define our own promises. In the next example we use the timeout function:

WinJS.Promise.timeout(3000).then(
    function(){console.log("done")},
    function(){console.log("fail")},
    function(){console.log("progress change")}
);

To be able to create our own promises we need to create our own instance of a Promise object
The constructor of a WinJS.Promise class accepts our three callbacks: success, fail and progress. In our implementation we need only to call our functions.

function someMethod() {
        return new WinJS.Promise(function (complete, error, progress) {
            // Some async calls
     progress(1);
     ...
            complete("okay");
        });
    }

In our case, the call to our class could look something like this:

var request = sommeMethod()
  .then(
   function() { console.log("done"); },
   function() { console.log("failed"); },
   function() { console.log("progress change");
});

For cancelation action we don't need any kind of custom think. We only need to call the cancel method. Automatically the error method will be called.

request.cancel();

When we are using promises we define a list of promises that we want to wait before the complete action is called. This action can be done using WinJS.Promise.join method. As the first parameter we need to define our promises. After that we can use then and define our callbacks.

WinJS.Promise.join([someMethod(), someMethod1("someParam")])
 .then(function() { console.log("done");});

If we want to end the call when one of the methods finish we would need to use WinJS.Promise.any.

WinJS.Promise.any([someMethod(), someMethod1("someParam")])
 .then(function() { console.log("done");});

When one of the calls finished, our complete method is automatically called, even if only one called is complete.
In this post we saw not only how we can use promises but how we can define them. The definition process of a promise is very simple and strait.
A post about Promises in JavaScript in a Metro Application: http://vunvulearadu.blogspot.com/2012/08/promises-and-asynchron-calls-in-metro.html