thoughts on coding

October 5, 2012

WinJS.UI.ListView and its LayoutNotInitialized exception

Filed under: Windows8, WWA — Tags: , — Frantisek @ 9:19 pm

While using WinJS.UI.ListView control and hiding/showing it together with binding/setting its datasource, it’s very possible to get the annoying exception “Layout is not initialized“.

Sample application

I developed a sampe app (after downloading just remove docx extension) where it’s possible to see when the exception is thrown.

The issue occurs in 2 cases:

  1. Given the ListView is hidden (display:none) and it is connected (via binding or JS) to empty list.
    When a new item is inserted into the datasource
    Then the ListView would like to render it but it’s necessary to measure the item size. Because of seeting “display:none” it is not possible, so mesureing fails and the exception is thrown
    Steps: check display none, click regenerate empty list, click add tile
  2. Given the ListView is hidden (display:none) and it is connected (via binding or JS) to non-empty list.
    When the items datasource is reset (via binding or JS) to some other non-empty datasource
    Then the ListView would like to render it but it’s necessary to measure the item size. Because of seeting “display:none” it is not possible, so mesureing fails and the exception is thrown
    Steps: check display none, click regenerate non-empty list


  1. using groupInfo property of layout which tells ListView: don’t measure on your own, here is the size
  2. when showing/hiding the listview we could enable/disable binding/setting the datasource

It’s possible to see both solution in the action in the downloaded application.


Solution a) is quite easy to do but it is not as performant as solution b.
Solution b) is better from the performance point of the view because when the listview is hidden and unbound, then it doesn’t receive and process any updates from datasource.

June 2, 2012

Observables and data-binding in action (Part III)

Filed under: Javascript, WWA — Tags: , , , — Frantisek @ 11:03 pm

Observables and their usage in binding

The typical scenario where observables’ features are leveraged is data-binding (binding the observable objects to html element – in other words model/viewmodel to view).

Declaration syntax

WinJS establish data-binding to UI in the following form:

<div data-win-bind="targetPropertyChain:sourcePropertyChain [customBindingFunction]" >

This was a declaration with the following meaning:

  • targetPropertyChain: declares a target of the databinding. It’s property name on the HTML element. Example: innerText. It supports also the chain of the properties, e.g. winControl.content.
  • sourcePropertyChain: declares a source of the databinding. It’s property name chain on an object (also called databinding context) whose data are used for binding. Example: address.street, name, etc.
  • customBindingFunction: it’s an optional parameter and declares a custom databinding function which replaces the default databinding function. There are several examples of it developed in MvvmJS. WinJS comes with two binding functions:
    1. default: it’s default oneway databinding function which binds a property from databinding context into a property on HTML element initialy and any time property on context object changes.
    2. onetime: it’s the function which bind the property from databinding context into element only once and only once.

Usage examples

<div data-win-bind="innerText:name" >

It’s the most simple binding use-case and declares binding from to targetElement.innerText.

<div data-win-bind="VISIBLE:isEmpty toggleClass" >

This defines the binding with the custom binding function defined as an extension in MvvmJS. The binding function toggleClass adds class VISIBLE to the element CSS class attribute in case datacontext.isEmpty is true.

<div data-win-bind="innerText:person.homeAddress.street" >

This defines binding street of the home address of the person accessible as a property on datacontext object to the innerText property on the target HTML element.

<div data-win-control="MvvmJS.UI.TemplateSelector" data-win-bind="winControl.content:person" >

First, it declares control MvvmJS.UI.TemplateSelector control (a custom WinJS UI control) over DIV. Then there is a databinding of datacontext.person to content property of winControl attribute set on the target DIV element. Note: it’s usual practice/pattern that UI control sets attribute winControl on the element to control instance.

Data-binding in action

That was just the first part of the story: declaration. Now, lets make the databinding alive.

WinJS introduces WinJS.Binding.processAll(root, dataContext, ..other parameters..) method which:

  1. searches for any element with data-win-bind attribute on the HTML element tree starting with root (optionaly skipping root) and
  2. then process the binding declarations.

WinJS.Binding.processAll method is usually called from JS page or fragment file.


(function () {
    "use strict";

    WinJS.UI.Pages.define("/views/dummyView.html", {
        element: null,

        ready: function (element, options) {
            this.element = element;
            element.winControl = this;
            this.options = options;

            var vm = new MvvmJS.Samples.DummyVieModel();
            WinJS.Binding.processAll(element, vm);

The example above is code-behind of the view DummyView which runs the binding processing in the ready event handler (raised when the page is rendered and ready to be used). In the ready event it instantiates DummyViewModel (please note a convention DummyView ~ DummyViewModel).

In case of an enterprise application it would be better to generalize the approach and define a base class with calling the binding process method from its ready method (this also brings MvvmJS).

Property chain subscriptions

There is one very usefull feature of WinJS databinding. WinJS databinding leverage observables and adds there property chain subscription of any property in sourcePropertyChain.

Let’s explain it on the example in BDD (behavior driven development) style:
Given: there is the following binding declaration

<div data-win-bind="innerText:person.address.street" >

When: person OR address OR street changes

Then: the binding is reprocessed and div’s innertext contains new value. This is very usefull feature in complex scenarios.

Comparing to manual approach

In case of manual approach it would be necessary to:

  1. initialy read the property and set the HTML element property with the value
  2. subscribe to any property chain change and handle the it with setting the new value

Manual approach implies more boilerplate and repetitive code which increase complexity of the result JS code.

In the next part we will look into custom binding method and extensions introduced in MvvmJS.

April 23, 2012

Testing WWA apps with Jasmine on Windows 8

Filed under: TDD, WWA — Tags: , , , , , , — Frantisek @ 9:51 pm

I work on a project which is WWA app and I must admit that:

  •  the code can quickly ends up in the mess,
  • the language itself is dynamic so there are no compile time errors (just runtime errors)

One of the ways how to ensure higher code quality is unit testing.

I paste here 2 vey important notes about unit testing from this blog because many people argue that they don’t have time to write unit tests OR it’s to advanced approach for them.

Testing is an advanced practice for the pros, not for me

My position is that whatever way you think of your process for writing code, you are testing it, e.g. by refreshing your browser(s) to verify that it is doing what it should. You are simply opting out of automating and improving your testing process, and in the long run (or even not-so-long run) you are spending more time hammering your browser’s refresh button than I spend writing tests that I can run today, tomorrow and next year should I so please.

Testing takes too much time, I’d rather just write production code

Both manual and automated testing takes time. But please, don’t spend an hour or two “evaluating” unit testing and/or TDD and then decide it is a waste of time. Unit testing and TDD are disciplines that need to be practiced just like any other. There is no way to get good at automated testing in a few hours. You need practice, and once you get there, then you will recognize the benefits I’m describing here, and then you will realize how much of a waste manual ad-hoc testing is. Besides, even if writing unit tests and testing your code rigorously takes a little more time, what would you prefer? To fail really fast, or to succeed?

Current situation

Unfortunatelly, there is no project template for unit testings of WWA apps in Visual Studio 2011. There are different projects templates for authoring WWA apps with VS 2011 but NO project template for unit testing and it’s a shame. We, inside Microsoft have the ways how to test WWA apps but they are not public.

The best would be to have a publicly available way how to test WWA apps with community behind.


First of all, there are some unit testing plugins into VS IDE like JustCode from Telerik or Unit testing plugin from JetBrains but no for running them as Metro WWA application.  So I had to invent my own solution. I had to port Jasmine runner including my own extensions in order to run the unit tests for WWA apps on Windows 8.

I use Jasmine as unit testing framework. I wrote a shell page (or so called test runner shell) which is set as the start page. The shell page runs the whole process which consist of:

  1. configuring Jasmine,
  2. discovering tests and
  3. running the tests upon user input.

In order to fit with Jasmine I had to write also a Jasmine reporter which communicates with the shell and reports the progress of tests execution.

Enought theory, example please.

I ported a referral example from Jasmine project and here is the result:

You can download the code from It’s in the folder misc\SampleTests

How all this works:

There are two main things:

  1. all test files ends with *tests.js. It’s configurable via jasmine.wwa.fileFilter. Example: jasmine.wwa.testFilesFilter = “*spec.js”;
  2.  unit testing project is configured to start test runner shell file: shell.html.

The shell loads and run ALL test files. A test file itself registers test suite and specs into the jasmine environment. Please note that shell runner follows M-V-VM style.

Future development

The shell runner will be updated quite much as MvvmJS is going to be extended with different aspects, e.g. binding extenions, controls, etc.

If you want to use it on your project, just copy the whole project and add your tests – the shell will discover your tests and you will be able to run them.

Happy testing on WWA!

Create a free website or blog at