<?xml version='1.0' encoding='UTF-8'?>
<page xmlns="http://projectmallard.org/1.0/"
xmlns:its="http://www.w3.org/2005/11/its"
type="topic"
id="image-viewer.js">
<info>
<title type="text">Image viewer (JavaScript)</title>
<link type="guide" xref="js#examples"/>
<desc>A little bit more than a simple "Hello world" application - write an image viewer in GTK+. Includes an introduction to the JavaScript language.</desc>
<revision pkgversion="0.1" version="0.1" date="2011-03-19" status="review"/>
<credit type="author">
<name>Jonh Wendell</name>
<email its:translate="no">jwendell@gnome.org</email>
</credit>
<credit type="author">
<name>Johannes Schmid</name>
<email its:translate="no">jhs@gnome.org</email>
</credit>
<credit type="editor">
<name>Marta Maria Casetti</name>
<email its:translate="no">mmcasettii@gmail.com</email>
<years>2013</years>
</credit>
</info>
<title>Image viewer</title>
<synopsis>
<p>In this tutorial, we're going to write a very simple GTK application that loads and displays an image file. You will learn how to:</p>
<list>
<item><p>Write a basic GTK user interface in JavaScript</p></item>
<item><p>Deal with events by connecting signals to signal handlers</p></item>
<item><p>Lay out GTK user interfaces using containers</p></item>
<item><p>Load and display image files</p></item>
</list>
<p>You'll need the following to be able to follow this tutorial:</p>
<list>
<item><p>An installed copy of the <link xref="getting-ready">Anjuta IDE</link></p></item>
<item><p>An installed copy of the <em>gjs</em> interpreter</p></item>
<item><p>Basic knowledge of any object-orientated programming language</p></item>
</list>
</synopsis>
<media type="image" mime="image/png" src="media/image-viewer.png"/>
<section id="anjuta">
<title>Create a project in Anjuta</title>
<p>Before you start coding, you'll need to set up a new project in Anjuta. This will create all of the files you need to build and run the code later on. It's also useful for keeping everything together.</p>
<steps>
<item>
<p>Start Anjuta and click <guiseq><gui>File</gui><gui>New</gui><gui>Project</gui></guiseq> to open the project wizard.</p>
</item>
<item>
<p>Choose <gui>Generic Javascript</gui> from the <gui>JS</gui> tab, click <gui>Forward</gui>, and fill out your details on the next few pages. Use <file>image-viewer</file> as project name and directory.</p>
</item>
<item>
<p>Click <gui>Apply</gui> and the project will be created for you. Open <file>src/main.js</file> from the <gui>Project</gui> or <gui>File</gui> tabs. It contains very basic example code.</p>
</item>
</steps>
</section>
<section id="js">
<title>JavaScript basics: Hello World</title>
<p>Before we start writing the image viewer, let's find out more about the way JavaScript is used in GNOME. Of course, your very first contact with any programming language should be the Hello World program which can already be found in <file>main.js</file>:</p>
<code mime="application/javascript">print ("Hello world!");</code>
<p>This should look quite natural if you're familiar with almost any other programming language. The function <code>print</code> is called with the argument <code>"Hello world!"</code>, which will be printed on the screen. Note that each line of code ends with a semicolon.</p>
</section>
<section id="classes">
<title>Classes in JavaScript</title>
<p>This is the standard way to define a class in JavaScript:</p>
<code mime="application/javascript" style="numbered"><![CDATA[
function MyClass () {
this._init ();
}
MyClass.prototype = {
_init: function () {
this.propertyA = "This is an object's field";
this.propertyB = 10;
},
aMethod: function (arg1, arg2) {
print ("inside aMethod: " + arg1 + " " + arg2);
},
dumpProperties: function () {
print (this.propertyA);
print (this.propertyB);
}
}]]></code>
<p>This defines a class called <code>MyClass</code>. Let's go through each part of the class definition:</p>
<steps>
<item>
<p><code>function MyClass</code> is the constructor of the class — its name must match the class's name. You can access any member of the class by using the <code>this</code> object; here, the constructor calls the class's <code>_init</code> method.</p>
</item>
<item>
<p>The <code>MyClass.prototype</code> block is where you define the <em>structure</em> of the class. Each class is made up of methods (functions) and fields (variables); there are three methods and two fields in this example.</p>
</item>
<item>
<p>The first method defined here is called <code>_init</code>, and we specify that it is a function with no arguments:</p>
<code>_init: function ()</code>
<p>We write the function inside some curly braces. Two fields are defined here, <code>propertyA</code> and <code>propertyB</code>. The first is set to a string and the second is set to an integer (10). The function doesn't return any value.</p>
</item>
<item>
<p>The next method is called <code>aMethod</code> and has two arguments, which it prints out when you call it. The final method is <code>dumpProperties</code>, and prints the fields <code>propertyA</code> and <code>propertyB</code>.</p>
</item>
<item>
<p>Note how the class definition (prototype) is arranged; each function definition is separated by a comma.</p>
</item>
</steps>
<p>Now that MyClass has been defined, we can play with it:</p>
<code mime="application/javascript" style="numbered"><![CDATA[
var o = new MyClass ();
o.aMethod ("Hello", "world");
o.propertyA = "Just changed its value!";
o.dumpProperties ();]]></code>
<p>This code creates a new instance of the class called <code>o</code>, runs <code>aMethod</code>, changes <code>propertyA</code> to a different string, and then calls <code>dumpProperties</code> (which outputs the fields).</p>
<p>Save the code in the <file>main.js</file> and then run it by using
<guiseq><gui>Run</gui><gui>Execute</gui></guiseq> from the menu or using the toolbar.</p>
</section>
<section id="gtk">
<title>A first Gtk application</title>
<p>Let's see what a very basic Gtk application looks like in JavaScript:</p>
<code mime="application/javascript" style="numbered"><![CDATA[
const Gtk = imports.gi.Gtk;
Gtk.init (null, null);
var w = new Gtk.Window ({title: "Image Viewer Demo"});
w.show ();
Gtk.main ();]]></code>
<p>Let's take a look at what's happening:</p>
<list>
<item>
<p>The first line imports the Gtk namespace (that is, it includes the Gtk library). The libraries are provided by GObject Introspection (gi), which provides language bindings for many GNOME libraries.</p>
</item>
<item>
<p><code>Gtk.init</code> initializes the Gtk library; this statement is mandatory for all Gtk programs.</p>
</item>
<item>
<p>The next line creates the main window by creating a new <code>Gtk.Window</code> object. You can pass several properties to the window's constructor by using the syntax <code>{property: value, property: value, ...}</code>. In this case we are setting the title of the window.</p></item>
<item><p>The next line explicitly shows the window. In Gtk, every widget is hidden by default.</p></item>
<item><p>Finally, <code>Gtk.main</code> runs the main loop — in other words, it executes the program. The main loop listens for events (signals) from the user interface and then calls a signal handler which will do something useful. We'll learn more about signals shortly.</p></item>
</list>
<p>Save the code in <file>main.js</file> and run it. You will notice that the application does not quit when you close the window. This is because we haven't set up a signal handler to deal with the window's <code>destroy</code> (close) signal yet. We'll do this shortly, but for now you can just hit <keyseq><key>Ctrl</key><key>C</key></keyseq> in the terminal window to quit the program.</p>
</section>
<section id="classes2">
<title>Adding classes</title>
<p>The proper way of doing Gtk programming is by using classes. Let's rewrite the simple code you just wrote using classes:</p>
<code mime="application/javascript" style="numbered"><![CDATA[
const Gtk = imports.gi.Gtk;
function ImageViewer () {
this._init ();
}
ImageViewer.prototype = {
_init: function () {
this.window = new Gtk.Window ({title: "Image Viewer Demo"});
this.window.show ();
}
}
Gtk.init (null, null);
var iv = new ImageViewer ();
Gtk.main ();]]></code>
<!-- FIXME: Throws an error, "JS ERROR: !!! Unhandled type int32 releasing GArgument" on Ubuntu 10.10 -->
<p>Notice that the program is the same; we just moved the window creation code to our own <code>ImageViewer</code> class. The class's constructor calls the <code>_init</code> method, which creates and shows the window. We then create an instance of the class before running the main loop (<code>Gtk.main</code>).</p>
<p>This code is modular and can be split into multiple files easily. This makes it cleaner and easier to read.</p>
</section>
<section id="signals">
<title>Signals</title>
<p>Signals are one of the key concepts in Gtk programming. Whenever something happens to an object, it emits a signal; for example, when a button is clicked it gives off the <code>clicked</code> signal. If you want your program to do something when that event occurs, you must connect a function (a "signal handler") to that signal. Here's an example:</p>
<code mime="application/javascript" style="numbered"><![CDATA[
function button_clicked () {
print ("you clicked me!");
}
var b = new Gtk.Button ({label:"Click me"});
b.connect ("clicked", button_clicked);]]></code>
<p>The last two lines create a <code>Gtk.Button</code> called <code>b</code> and connect its <code>clicked</code> signal to the <code>button_clicked</code> function, which is defined above. Every time the button is clicked, the code in the <code>button_clicked</code> function will be executed. It just prints a message here.</p>
<p>The syntax for connecting any signal to a function is:</p>
<code mime="application/javascript"><![CDATA[
object.connect (<signal_name>, <function_to_be_called>);]]></code>
<p>You can find signal definitions for any object in the <link href="https://developer.gnome.org/gtk3/stable/gtkobjects.html">GTK class reference</link>.</p>
<note>
<p>You can simplify the code by making use of an inline function definition:</p>
<code mime="application/javascript"><![CDATA[
b.connect ("clicked", function () { print ("you clicked me!"); });]]></code>
</note>
</section>
<section id="close">
<title>Closing the window</title>
<p>When you close a Gtk window it's not really closed, it's hidden. This allows you to keep the window around (which is useful if you want to ask the user if they really want to close the window, for example).</p>
<p>In our case, we really do just want to close the window. The simplest way of doing this is by connecting the <code>hide</code> signal of the GtkWindow object to a function that closes the application. Go back to the <file>image-viewer.js</file> file and add the following code to the <code>_init</code> method, on the line above <code>this.window.show</code>:</p>
<code mime="application/javascript" style="numbered">this.window.connect ("hide", Gtk.main_quit);</code>
<p>This connects the <code>hide</code> signal of the window to Gtk's <code>main_quit</code> function, which ends the execution of the Gtk main loop. Once the main loop finishes, the function <code>Gtk.main</code> returns. Our program would continue to run any code written after the <code>Gtk.main ();</code> line, but since we don't have any code after that point, the program just ends.</p>
</section>
<section id="containers2">
<title>Containers: Laying-out the user interface</title>
<p>Widgets (controls, such as buttons and labels) can be arranged in the window by making use of <em>containers</em>. You can organize the layout by mixing different types of containers, like boxes and grids.</p>
<p>A <code>Gtk.Window</code> is itself a type of container, but you can only put one widget directly into it. We would like to have two widgets, an image and a button, so we must put a "higher-capacity" container inside the window to hold the other widgets. A number of <link href="http://library.gnome.org/devel/gtk/stable/GtkContainer.html">container types</link> are available, but we will use a <code>Gtk.Box</code> here. A <code>Gtk.Box</code> can hold several widgets, organized horizontally or vertically. You can do more complicated layouts by putting several boxes inside another box and so on.</p>
<note>
<p>There is a graphical user interface designer called <app>Glade</app> integrated in <app>Anjuta</app> which makes UI design really easy. For this simple example, however, we will code everything manually.</p>
</note>
<p>Let's add the box and widgets to the window. Insert the following code into the <code>_init</code> method, immediately above the <code>this.window.show</code> line:</p>
<code mime="application/javascript" style="numbered"><![CDATA[
var main_box = new Gtk.Box ({orientation: Gtk.Orientation.VERTICAL, spacing: 0});
this.window.add (main_box);]]></code>
<p>The first line creates a <code>Gtk.Box</code> called <code>main_box</code> and sets two of its properties: the <code>orientation</code> is set to vertical (so widgets are arranged in a column), and the <code>spacing</code> between the widgets is set to 0 pixels. The next line then adds the newly-created <code>Gtk.Box</code> to the window.</p>
<p>So far the window only contains an empty <code>Gtk.Box</code>, and if you run the program now you will see no changes at all (the <code>Gtk.Box</code> is a transparent container, so you can't see that it's there).</p>
</section>
<section id="packing2">
<title>Packing: Adding widgets to the container</title>
<p>To add some widgets to the <code>Gtk.Box</code>, insert the following code directly below the <code>this.window.add (main_box)</code> line:</p>
<code mime="application/javascript" style="numbered"><![CDATA[
this.image = new Gtk.Image ();
main_box.pack_start (this.image, true, true, 0);]]></code>
<p>The first line creates a new <code>Gtk.Image</code> called <code>image</code>, which will be used to display an image file. Then, the image widget is added (<em>packed</em>) into the <code>main_box</code> container using <code>Gtk.Box</code>'s <link href="http://library.gnome.org/devel/gtk/stable/GtkBox.html#gtk-box-pack-start"><code>pack_start</code></link> method.</p>
<p><code>pack_start</code> takes 4 arguments: the widget that is to be added to the <code>Gtk.Box</code> (<code>child</code>); whether the <code>Gtk.Box</code> should grow larger when the new widget is added (<code>expand</code>); whether the new widget should take up all of the extra space created if the <code>Gtk.Box</code> gets bigger (<code>fill</code>); and how much space there should be, in pixels, between the widget and its neighbors inside the <code>Gtk.Box</code> (<code>padding</code>).</p>
<p>Gtk containers (and widgets) dynamically expand to fill the available space, if you let them. You don't position widgets by giving them a precise x,y-coordinate location in the window; rather, they are positioned relative to one another. This makes handling window resizing much easier, and widgets should automatically take a sensible size in most situations.</p>
<p>Also note how the widgets are organized in a hierarchy. Once packed in the <code>Gtk.Box</code>, the <code>Gtk.Image</code> is considered a <em>child</em> of the <code>Gtk.Box</code>. This allows you to treat all of the children of a widget as a group; for example, you could hide the <code>Gtk.Box</code>, which would also hide all of its children at the same time.</p>
<p>Now insert these two lines, below the two you just added:</p>
<code mime="application/javascript" style="numbered"><![CDATA[
var open_button = new Gtk.Button ({label: "Open a picture..."});
main_box.pack_start (open_button, false, false, 0);]]></code>
<p>These lines are similar to the first two, but this time they create a <code>Gtk.Button</code> and add it to <code>main_box</code>. Notice that we are setting the <code>expand</code> argument (the second one) to <code>false</code> here, whereas it was set to <code>true</code> for the <code>Gtk.Image</code>. This will cause the image to take up all available space and the button to take only the space it needs. When you maximize the window, the button size will remain the same, but the image size will increase, taking up all of the rest of the window.</p>
<p>Finally, we must change the <code>this.window.show ();</code> line to read:</p>
<code>this.window.show_all ();</code>
<p>This will show the child of the Gtk window, and all of its children, and its children's children, and so on. (Remember that Gtk widgets are all hidden by default.)</p>
</section>
<section id="loading2">
<title>Loading the image: Connecting to the button's <code>clicked</code> signal</title>
<p>When the user clicks on the <gui>Open</gui> button, a dialog should appear so that the user can choose a picture. Once chosen, the picture should be loaded and shown in the image widget.</p>
<p>The first step is to connect the <code>clicked</code> signal of the button to a signal handler function, which we call <code>_openClicked</code>. Put this code immediately after the <code>var open_button = new Gtk.Button</code> line where the button was created:</p>
<code mime="application/javascript"><![CDATA[
open_button.connect ("clicked", Lang.bind (this, this._openClicked));]]></code>
<p>We are using the <em>Lang</em> JavaScript helper here. It allows us to connect a <em>class method</em> to the signal, rather than a plain function (without a class) which we had used before for the window's <code>hide</code> signal. Don't worry about this for now, it's just a technical detail. For it to work, you also need to put the following line at the top of the file:</p>
<code mime="application/javascript">const Lang = imports.lang;</code>
</section>
<section id="loading3">
<title>Loading the image: Writing the signal's callback</title>
<p>Now we can create the <code>_openClicked()</code> method. Insert the following into the <code>ImageViewer.prototype</code> code block, after the <code>_init</code> method (and not forgetting the comma):</p>
<code mime="application/javascript" style="numbered"><![CDATA[
_openClicked: function () {
var chooser = new Gtk.FileChooserDialog ({title: "Select an image",
action: Gtk.FileChooserAction.OPEN,
transient_for: this.window,
modal: true});
chooser.add_button (Gtk.STOCK_CANCEL, 0);
chooser.add_button (Gtk.STOCK_OPEN, 1);
chooser.set_default_response (1);
var filter = new Gtk.FileFilter ();
filter.add_pixbuf_formats ();
chooser.filter = filter;
if (chooser.run () == 1)
this.image.file = chooser.get_filename ();
chooser.destroy ();
}]]></code>
<p>This is a bit more complicated than anything we've attempted so far, so let's break it down:</p>
<list>
<item>
<p>The line beginning with <code>var chooser</code> creates an <gui>Open</gui> dialog, which the user can use to choose files. We set four properties: the title of the dialog; the action (type) of the dialog (it's an "open" dialog, but we could have used <code>SAVE</code> if the intention was to save a file; <code>transient_for</code>, which sets the parent window of the dialog; and <code>modal</code> which, if set to <code>true</code>, prevents the user from clicking on another area of the application until the dialog is closed.</p>
</item>
<item>
<p>The next two lines add <gui>Cancel</gui> and <gui>Open</gui> buttons to the dialog. The second argument of the <code>add_button</code> method is the (integer) value that is returned when the button is pressed: 0 for <gui>Cancel</gui> and 1 for <gui>Open</gui>.</p>
<p>Notice that we are using <em>stock</em> button names from Gtk, instead of manually typing "Cancel" or "Open". The advantage of using stock names is that the button labels will already be translated into the user's language.</p>
</item>
<item>
<p><code>set_default_response</code> determines the button that will be activated if the user double-clicks a file or presses <key>Enter</key>. In our case, we are using the <gui>Open</gui> button as default (which has the value 1).</p>
</item>
<item>
<p>The next three lines restrict the <gui>Open</gui> dialog to only display files which can be opened by <code>Gtk.Image</code>. A filter object is created first; we then add all kinds of files supported by <code>Gdk.Pixbuf</code> (which includes most image formats like PNG and JPEG) to the filter. Finally, we set this filter to be the <gui>Open</gui> dialog's filter.</p>
</item>
<item>
<p><code>chooser.run</code> displays the <gui>Open</gui> dialog. The dialog will wait for the user to choose an image; when they do, <code>chooser.run</code> will return the value <output>1</output> (it would return <output>0</output> if the user clicked <gui>Cancel</gui>). The <code>if</code> statement tests for this.</p>
</item>
<item><p>Assuming that the user did click <gui>Open</gui>, the next line sets the <code>file</code> property of the <code>Gtk.Image</code> to the filename of the image selected by the user. The <code>Gtk.Image</code> will then load and display the chosen image.</p>
</item>
<item>
<p>In the final line of this method, we destroy the <gui>Open</gui> dialog because we don't need it any more.</p>
</item>
</list>
</section>
<section id="run">
<title>Run the application</title>
<p>All of the code you need should now be in place, so try running the code. That should be it; a fully-functioning image viewer (and a whistlestop tour of JavaScript and Gtk) in not much time at all!</p>
</section>
<section id="impl">
<title>Reference Implementation</title>
<p>If you run into problems with the tutorial, compare your code with this <link href="image-viewer/image-viewer.js">reference code</link>.</p>
</section>
<section id="next">
<title>Next steps</title>
<p>Here are some ideas for how you can extend this simple demonstration:</p>
<list>
<item>
<p>Have the user select a directory rather than a file, and provide controls to cycle through all of the images in a directory.</p>
</item>
<item>
<p>Apply random filters and effects to the image when it is loaded and allow the user to save the modified image.</p>
<p><link href="http://www.gegl.org/api.html">GEGL</link> provides powerful image manipulation capabilities.</p>
</item>
<item>
<p>Allow the user to load images from network shares, scanners, and other more complicated sources.</p>
<p>You can use <link href="http://library.gnome.org/devel/gio/unstable/">GIO</link> to handle network file transfers and the like, and <link href="http://library.gnome.org/devel/gnome-scan/unstable/">GNOME Scan</link> to handle scanning.</p>
</item>
</list>
</section>
</page>