Organizing your WinJS Metro code into namespaces

I have been struggling with namespacing since I started with WinJS. Not sure why, just can never get the Magical Rites correct. I think is has to do with when to add the Feral Cat Whiskers.

Anyway, here is a list of how I made it work for a simple reference project.

1) Add a base.js file to your /js directory. Populate it with a namespace declaration. DO NOT FORGET the closing parens.

(function () {
    WinJS.Namespace.define('MyApp', {});
})();

2) Add a settings.js file to the /ja folder (for example). Populate it with a method to share with the namespace.

(function () {
    function getSettings() {
        var settings = "D";
    }
    WinJS.Namespace.defineWithParent(MyApp, 'Settings', {
        getSettings: getSettings,
    });
})();

3) In the default.html file, add a reference to the two new files.

<!DOCTYPE html>
<html>
<head>
    <meta charset="utf-8" />
    <title>Namespace</title>

    <!-- WinJS references -->
    <link href="//Microsoft.WinJS.1.0.RC/css/ui-dark.css" rel="stylesheet" />
    <script src="//Microsoft.WinJS.1.0.RC/js/base.js"></script>
    <script src="//Microsoft.WinJS.1.0.RC/js/ui.js"></script>

    <!-- Namespace references -->
    <link href="/css/default.css" rel="stylesheet" />
    <script src="/js/default.js"></script>
    <script type="text/javascript" src="js/base.js"></script>
    <script type="text/javascript" src="js/settings.js"></script>
</head>
<body>
    <p>Content goes here</p>
</body>
</html>

4) In default.js, call MyApp.Settings.getSettings() in the onactivated event. It's the bold line below.

// For an introduction to the Blank template, see the following documentation:
// http://go.microsoft.com/fwlink/?LinkId=232509
(function () {
    "use strict";

    var app = WinJS.Application;
    var activation = Windows.ApplicationModel.Activation;
    WinJS.strictProcessing();

    app.onactivated = function (args) {
        if (args.detail.kind === activation.ActivationKind.launch) {
            if (args.detail.previousExecutionState !== activation.ApplicationExecutionState.terminated) {
                MyApp.Settings.getSettings();
            } else {
                // TODO: This application has been reactivated from suspension.
                // Restore application state here.
            }
            args.setPromise(WinJS.UI.processAll());
        }
    };

    app.oncheckpoint = function (args) {
        // TODO: This application is about to be suspended. Save any state
        // that needs to persist across suspensions here. You might use the
        // WinJS.Application.sessionState object, which is automatically
        // saved and restored across suspension. If you need to complete an
        // asynchronous operation before your application is suspended, call
        // args.setPromise().
    };
    
    app.start();
})();

That's it. All there is too it. Everything else is gravy. At least this will give me something to go back and look at when I forget how to do it.

Make sure to get my new Windows 8 Programming with HTML5 For Dummies book when it comes out.

Doing security analysis on Windows 8 Metro Apps with Zed Attack Proxy

This is the first in a series of articles about attacking Windows 8 applications using Zed Attack Proxy, or ZAP. Windows 8 is the new version of the venerable Windows operating system from Microsoft. It has a new Windows programming model that is heavily based in a new API for Windows development, called WinRT. ZAP is a web proxy for security analysis of applications by OWASP. It is free.

Setting up ZAP for Windows 8

After installing ZAP, it needs to be made ready to act as a proxy for Metro applications. As most Metro applications make heavy use of the Internet for data storage, information access and communication with the Windows Store, use of a proxy is a great place to start to test the underlying security.

1) Press the Windows button and type Internet Options.

2) Press the Settings tab on the right of the Start Screen.

3) Select the Internet Options control panel.

4) Click the Connections tab.

5) Click the LAN Settings button

6) Change the Provy Server settings like Figure 1. Check the 'Use a proxy server' checkbox, then set the address to 'localhost' and the port to '8080'.

 

Figure 1 - configuring a proxy

7) Click OK, then click OK.

Finally, you have to configure the app you are testing to allow the loopback (locakhost) address as a proxy, which by default AppContainers aren't allowed to do. You can do this with PowerShell, but it is a lot easier to download Eric Law's excellent EnableLoopback utility, which will do the work for you. You just need to find the app you want to test and select the checkbox, then click Save Changes, like Figure 2.

 

Figure 2: the EnableLoopback Utility

Not your machine is configured to use ZAP as a proxy. While these settings are set, you will have to have ZAP running to use the Internet. Also, enabling the loopback address circumvents important security controls placed on Metro apps. Only use it on a test system.

Testing your settings

To test your settings, run ZAP by pressing the Windows key and typing ZAP then pressing enter. Then run an installed Metro app, like the Finance app, which you have configured for loopback. In ZAP, all of the services being called by the application will appear in the Sites pane, and all of the individual calls will appear in the History tab. Figure 3 shows my results.

 

Figure 3 - Finance Test results

 Each of these represents an HTTP call being made by the Finance app to populate it's screens. Apparently, the developers of this app chose not to use SSL for any of the calls, which makes our job easier - although since we have a man in the  middle the analysis can still be carried forward. Either way, this gives us a ripe field to begin our analysis.

Attacking underlying services

You can now use ZAP to test the underlying services with the fuzzer.

1) Select a service to test in the Sites pane. I selected the api.bing.com/appex service.

2) Highlight a parameter to fuzz in the Request pane, right in the querystring. I selected the 'a' parameter in Figure 4.

 

Figure 4 - Setting the fuzzing dictionary

3) Right click and select Fuzz. The dialog that appears is also shown in Figure 4.

4) Select the Fuzz category in the dropdown. I selected SQL Injection in Figure 4.The list of pre-installed fuzzers appear in the Fuzzers listbox. These are known attacks that work against a variety of databases. As this is a Microsoft service, we can assume that MSSQL is probably the database, although other test tools could be used to ascertain this. In Figure 4 that is the fuzzer I selected.

5) Click Fuzz to begin your attack. ZAP will begin to call the service with the attacks in the SQL Injection file.

And ... it didn't work at all. The Fuzzer tab shows each attack, and all of them have a 404 error in the result column. Apparently that field is being used as a controller method name - I should have thought of that!

Oh wrell, let's try something else. Select the 'q' parameter and try again with a SQL attack. This one is a lot better - every result comes back as a 200. Awesome.  Click on a line in the Fuzzer tab to see the result in the Response tab.

{
   "BdiGeneric_BingResponse_1_0":{
      "AppNS":"AppEx",
      "Responses":[

      ],
      "Errors":[
         {
            "Code":1002,
            "Message":"Parameter has invalid value.",
            "Parameter":"AppExNewsVerticalRequest.AppId",
            "Value":"'create user name identified by pass123 temporary tablespace temp default tablespace users; "
         }
      ]
   }
}

What have we here? A error code 1002? Aah, 'Parameter has invalid value.' -- they are using parameterized queries. I suppose one would expect a well vetted service like Bing to be hardened against the #1 exploit on the OWASP Top 10 though.

In the next in this series we will look at an exploitable app and try and test the services that make up its data storage facility. Thanks for reading!

Geting Notified

This is an excerpt from the upcoming book Programming Windows 8 with HTML5 for Dummies

--------------------------------------

 Now that you have the template set up, getting notifications on the tile is not that tough. More or less, you want to decide what events in your system should update the tile, and then attach tile updates to them.

 So think about your application. In the ToDoToday app, I update the tile whenever the user updates the task list, making sure I show the latest tasks in the m notification area. If writing a game, you probably would want to update the tile every time the user clears a level.

 In general, treat the notification as exactly what it is. It is the opportunity to notify the user what the last thing happening in the app is. So:

 Update the notifications when the user does something significant

  • Update the notifications when there is significant information to pass on to the user.
  • Update the notifications when the app is about to be suspended.

 

Making notifications part of your tile

 The good news is that you have already made a notification part of your tile, in Rendering Content. Setting the Text element value in the tile XML is essentially a notification - the only difference is in use and strategy - the tactics are the same.

 So to best implement it probably would be best to construct a function that is designed to be called from any of the important events.

 

function updateTile(textNotification) {
    var tileTemplate = Windows.UI.Notifications.TileUpdateManager.getTemplateContent(Windows.UI.Notifications.TileTemplateType.tileWideImageAndText01);
    var tileTextAttributes = tileTemplate.getElementsByTagName("text");
       tileTextAttributes[0].appendChild(tileTemplate.createTextNode(textNotification));
    var tileImageAttributes = tileTemplate.getElementsByTagName("image");
    tileImageAttributes[0].setAttribute("src", "ms-appx:///images/WideLogo.png");
     var tileNotification = new Windows.UI.Notifications.TileNotification(tileTemplate);
    Windows.UI.Notifications.TileUpdateManager.createTileUpdaterForApplication().update(tileNotification);
}

 

This we can then call, and maybe pass in the date just to see how it works. For instance, add a promise activated function after the processAll function in the default.js file of the sample project.

 

WinJS.UI.processAll().done(function (e) {
    var today = new Date();
    updateTile(today);
});

 

Run this and the tile should be updated with the date in the text region. Now, let’s say we want to update the tile live as the user is using the app. To replicate this, you can add a button and an event handler for said button right there on the default page.

 

<body>
   <button type="button" id="actionButton" style="position: relative; left: 300px; top: 300px;">Click Me</button>
</body>

 

Set up a variable for the button in default.js (I called it theButton) and then add a little code to call the updateTile function:

 

theButton = document.getElementById("actionButton")
theButton.addEventListener("click", function (e) {
    var today = new Date();
    updateTile(today);
});

 

Run the app again, then press the Windows key and confirm the date and time is in the tile. Then go back to the sample app and click the button.

 

Press the Windows key again and the tile will have changed. This could be the score, the news, or an appointment. The choice is yours.

 

Getting a message on the queue

 

What if there is more than one thing you want to show the user? You could use one of the templates with a lot of small text, but there is a better choice. Microsoft allows you to stack up to 5 messages on the queue and the tile will cycle through them one at a time.

 

In order to do this you MUST enable the queue. This is a one line property setting that you can put in the initialization code of your app. I used the onactivated event for this example:

 

Windows.UI.Notifications.TileUpdateManager.createTileUpdaterForApplication().enableNotificationQueue(true);

 

After you do that, I would recommend clearing the queue. You don’t want what was in there before rolling through with your new content.

 

Windows.UI.Notifications.TileUpdateManager.createTileUpdaterForApplication().clear();

 

So now our onAppActivated looks like this:

 

app.onactivated = function (eventObject) {
    if (eventObject.detail.kind === Windows.ApplicationModel.Activation.ActivationKind.launch) {
        if (eventObject.detail.previousExecutionState !== Windows.ApplicationModel.Activation.ApplicationExecutionState.terminated) {
    Windows.UI.Notifications.TileUpdateManager.createTileUpdaterForApplication().enableNotificationQueue(true);
Windows.UI.Notifications.TileUpdateManager.createTileUpdaterForApplication().clear();
            theButton = document.getElementById("actionButton")
            theButton.addEventListener("click", function (e) {
               var today = new Date();
               updateTile(today);
            });
        }
        WinJS.UI.processAll().done(function (e) {
            var today = new Date();
            updateTile(today);
        });
    }
};

Press F5 to start the app, and then click the Windows button to take a look at the Tile. It should have the date on it. Go get a cup of coffee, and come back, I’ll wait.

 

Now press the Click Me button and then take a look at the tile. The notification should be scrolling between the two times - when you started the app, and when you clicked the button.

 

Go back to the app, click the button again. Now look at the tile; there are three dates scrolling by. You can have up to five of these, and the Notification manager will drop the oldest automatically.

 

Animating Windows

 

This is an excerpt from the upcoming book Programming Windows 8 with HTML5 for Dummies

-----------------------------------------

The component animation that is inherent to HTML5 and even JQuery is not only supported in Metro, it is encouraged. The navigation model is new through, and there are bits in WinRT that make it easy to animate.

 

Everything for page and content level animation (sometimes called Transitions) is covered by WinJS though, and you can find it in WinJS.UI.Animation. Here is a sample of what you can get there:

 

  • Page transition: Animates the contents of a page into or out of view.
  • Content transition: Animates one piece or set of content into or out of view.
  • Fade in/out: Shows transient elements or controls.
  • Crossfade: Refreshes a content area.
  • Expand/Collapse: Shows additional inline information.
  • Reposition: Moves an element into a new position.
  • Show/Hide popup: Displays contextual UI on top of the view.
  • Show/Hide edge UI: Slides edge-based UI into or out of view.
  • Show/Hide panel: Slides large edge-based panels into or out of view.
  • Add/Delete from list: Adds or deletes an item from a list.
  • Add/Delete from search list: Adds or deletes an item from a list when filtering search results.
  • Badge update: Updates a numerical badge.
  • Start/End a drag or drag-between: Gives visual feedback during a drag-and-drop operation.

 

That’s a lot of animating. Don’t feel like you have to animate everything to get into the store. I would, though, seriously consider animating page and content transitions in order to show that you have given a nod to the realities of the Metro design language.

 

Page level transitions

 

Compared to dealing with the GestureLibrary, page transitions are a snap. In WinJS.UI.Animation you will find an enterPage function, and an exitPage function. They both do about the same thing, they just need to be run at different times in the page lifecycle.

 

Both methods take an element object and an offset object, and return a Promise. While an element is in transition you can’t act on it, so you want to use the promise to delay any page functionality (like removing an element) until the animation is over.

 

WinJS.UI.Animation.enterPage(content, startingPosition).done();

 

The content is the element or elements that you are animating. This seems counterintuitive but I believe it is set up this way so that you can choose to animate only certain parts of the page, and still have content level control later. The enterPage function is designed to be used with the Navigation libraries we talked about in Saying Hello World.

 

The startingPosition is exactly what it looks like. It is where the content collection should start before the animation occurs. You can pick a place on the page and pass in a JavaScript array of the location to make this happen. The recommendation is, though, to use “undefined” for the default start location.

 

The done of course represents the Promise that is being returned by the function. When the animation is done, the rest of the code will be run, and not until the animations is done. This is to prevent the entire UI from locking up while an animation is being performed.

 

To set up an animated page sequence, try the following steps:

 

  • Create a blank JavaScript Metro project.
  • Make a new directory called “pages” for the pages we are going to transition amongst.
  • Add three pages. Right click on the pages directory and select Add New Item. Select Page Control. Do it three times. I named mine firstpage, nextpage and thepageafter.
  • In default.html, add a content div replacing the paragraph tag, like this:

 

<body>
    <div id="content"></div>
</body>

 

  • In the default.js file, add a navigated event handler.

 

WinJS.Navigation.addEventListener("navigated", function (eventObject) {
        var url = eventObject.detail.location;
        var content = document.getElementById("content");
        WinJS.Utilities.empty(content);
        eventObject.detail.setPromise(WinJS.UI.Pages.render(url, content, eventObject.detail.state).then(function () {
            WinJS.Application.sessionState.lastUrl = url;
        }));
    });

 

  • Also, navigate to the first content page in the onactivated event. This will get things started.

 

app.onactivated = function (args) {
        if (args.detail.kind === activation.ActivationKind.launch) {
            args.setPromise(WinJS.UI.processAll().then(function(){
                return WinJS.Navigation.navigate("/pages/firstpage.html");
            }));
        }
    };

 

  • The firstpage.html, nextpage.html, and thepageafter.html give the header and main sections a ID. While you are in there, add a button for navigation.

 

<body>
    <div class="firstpage fragment">
        <header aria-label="Header content" role="banner" id="header">
            <button class="win-backbutton" aria-label="Back" disabled></button>
            <h1 class="titlearea win-type-ellipsis">
                <span class="pagetitle">Welcome to firstpage</span>
            </h1>
        </header>
        <section aria-label="Main content" role="main" id="main">
            <p>Content goes here.</p>
            <button id="nextPage">Next Page</button>
        </section>
    </div>
</body>

 

Add a transition function to firstpage.js, nextpage.js and thepageafter.js that navigates to the shosen page and uses the animation.

 

function transition() {
        WinJS.UI.Animation.exitPage([[header],[main]], null).done(
            function () {
                WinJS.Navigation.navigate("/pages/nextpage.html");
            });
    }

 

This will be different for each page! It should navigate to the page you want to go next.

 

  • Finally, in the firstpage.js (and the others) and change the ready event to fire the new transition function when the click event occurs.

 

WinJS.UI.Pages.define("/pages/firstpage.html", {
        // This function is called whenever a user navigates to this page. It
        // populates the page elements with the app's data.
        ready: function (element, options) {
            var nextPage = document.getElementById("nextPage");
            nextPage.addEventListener("click", transition, false);
            WinJS.UI.Animation.enterPage([[header],[main]], null);
        }
    });

I wish I had a pretty picture of this to show you, but wiley STILL hasn’t implemented those animated pages. If you can’t get it to work, download the sample code and give that a try. There is also a fantastic sample (if a little overcomplicated) on the dev.windows.com samples site.

 

Animating content

 

Good news! Animating content is almost excactly the same as animating entire pages. The Animation library has enterContent and exitContent functions that work a lot like enterPage and exitPage.

 

WinJS.UI.Animation.enterContent(content, startingPosition).done()

 

The content object is an HTML element (the ID of said object) that you want to animate. The startingPosition is best set to null unless you have a specific starting position. And the Promise returned is just to make sure that you don’t do anything with the entity until the animation is through.

 

I’ll spare you the steplist – it looks a lot like the Animating Pages example above. Just put the enterContent in the load function, and the exitContent in the navigation event.

 

Treating users right

This is an excerpt from the upcoming book Programming Windows 8 with HTML5 for Dummies

----------------------------------------

Metro apps are different.  As I mentioned in This Is Not .NET, Metro apps are optimized for touch, usually mashups, and often single use. They are not complex, menu driven or multi-functional.  They are usually for consumption, not creation. It is a totally different user experience that we are after here.

Because of this, there is a totally different design paradigm.  Now, Microsoft has tried to lay on the User Interface guidelines before with mixed success, but this time they are serious! The development environment is set up to only allow certain kinds of user interface elements, and if you step away from the path you will find your going rather tough.

Setting some design principles

Microsoft is being very clear in laying out some user interface and architectural design patterns.  It remains to be seen if these will be upheld in the review process, but they are good guidelines and should be followed.

While we will go into more detail in Chapter 2, there are a few core principles I want to use to describe Metro apps that will make the rest of this chapter make more sense. Metro apps are:

  • Socially connected. Metro apps make use of social network and public sharing networks to broadcast information, connect with friends, and make new groups.
  • Living in a sandbox. An install of the Metro app should never alter the core functionality of the user’s machine.
  • Designed for touch. These apps are designed to be run on a tablet, but should work with a mouse. Much more on this later.
  • Screen size sensitive. Like a web page, Metro apps should be sensitive to resolution, orientation, and docking.
  • Made for multitasking. Users can ‘snap’ Metro apps into specified parts of the screen, and if you want a nice looking app you have to handle that in your interface.

Using the Design Checklist

Something that Microsoft has implemented to help with the core principles is the Design Checklist. It covers what you need to make sure you have planned for in your app.

                           *  Basic design groups – core things you need to make sure you have covered

                                       *App tile – the icon that the user touches to launch your app

                                       *Splash screen – what the user sees while your app is loading

                                       *Application Lifecycle – be ready when the app gets suspended, or the battery runs out.

                                       *App bar – all your apps commands go in the same place.

                           *  Engaging the user with app contracts

                                       *Settings – how the app communicates with Windows

                                       *Search – how does it apply to your app?

                                       *Share – be social!

                                       *Play to – stream to other devices

                           *  Various Views

                                       *Full Screen

                                       *Snapped

                                       *Filled

                                       *Portrait

                                       *Scaled

                           *  Adding features – take advantage of the neat stuff you can do

                                       *Semantic Zoom

                                       *Notifications

                                       *Roaming

                                       *Content tiles

                                       *Gesture library

                                       *File picker

                                       *Animations

                                       *User tiles

Convention over configuration

You may have noticed by now that the Metro style is a lot about convention – convention being defined as how Apple decided to make the IPad work. You don’t actually have a lot of artistic license over how the user will interact with the meta-information related to your app; it is pretty well defined by the Design Checklist and Core Principles.

Let’s take navigation for example.  In Windows Forms application, navigation is all over the place.  There are menus and ribbons and buttons (oh my!). In MetroUI, however, the navigation of the application should be document oriented, and in the app bar, at the bottom of the application.

The programming model supports this, in fact.  It supports it so well that there really isn’t any other way to implement the app bar.  To see what I mean, open up the Hello World application we made in Getting There Ourselves.

In the HTML, inside the body tag, add a div that represents the AppBar control.  Notice the data-win-control declaration:

<body>

    <h1 id="headline"></h1>

   <div data-win-control="WinJS.UI.AppBar" id="appbar">

   <button data-win-control="WinJS.UI.AppBarCommand" data-win-options="{icon:'back', id:'', label:'example', onclick:null, section:'global', type:'button'}"></button>

   </div>

</body>

Now you can run the application, and have a nice back button in the app bar. All of your navigational type stuff would go here – start over, change views, go home.  Not in a menu, not in the page.  In the app bar.

CSSLP continued - uCertify's web client

Well, I had planned to take my CSSLP in April but life got in the way, so I am shooting for more of an August timeframe now, after the book is written.

uCertify has a web client that I thought would work out well for me since I keep moving back and forth between different machines. I think they are pushing to the web to draw in tablet users, but either way, it works out well for me.

The web is a lot like the PC version in functionality. They have four test styles - the pre-assessment, the quiz, a regular practice test, and a new thing, Flash Cards, which I think are very cool. There are four study materials, too: the textbook, reports on your trials, a planner and the Final Test, which you should get a 95 on before you go take the exam.

The whole thing is implemented in HTML, and is occasionally a little choppy. I am running it in Firefox right now, and it is a little better than it behaved in IE10. It is very touch friendly and runs well on my Windows 8 slate.

The content is just as good as the PC version, and I feel very secure having it as my study partner. The flash cards are awesome - just like something you might have written for yourself in school: the term on one side, the definition on the other. I think this plus the official textbook will get me through the test. We'll see in August, though. The proof is in the pudding!

You can learn more about the uCertify web client at learn.ucertify.com.

Bill Sempf

Husband. Father. Pentester. Secure software composer. Brewer. Lockpicker. Ninja. Insurrectionist. Lumberjack. All words that have been used to describe me recently. I help people write more secure software.

PageList

profile for Bill Sempf on Stack Exchange, a network of free, community-driven Q&A sites

MonthList