Thursday, February 21, 2013

Dynamically resizing Google Chrome Extensions

Lately I've been working on an extension for Google Chrome to read the last entry of the famous xkcd comic. At a certain point I stopped because I didn't understand how to dynamically resize the popup window of the Chrome extension. This required quite some research to understand what was going wrong, here is how I created the extension, which problem I had and how I solved it.

How I did it

The extension is very easy, it executes a GET request to the xkcd website, gets the content of the index page and returns it. What I used to do, for a matter of simplicity, was getting the content, paste it in an hidden div, get the image of the comic of the day (through a getElementById call) and put it in a visible div.

The Problem

What one would expect is that the Google Chrome popup page of the extension notice the change of content and resize itself to fit the new content. But this doesn't happen. The popup page remains as small as it can get without styles, or of a fixed with/height if specified in the css file. I didn't wanted to have a fixed width/height as comics usually have not a fixed size (sometimes they are as small as a strip, sometimes they are quite big).
Since the popup extension doesn't resize, I though about resizing it using JavaScript:

var width =  document.getElementById("comic").offsetWidth + 10;
var height = document.getElementById("comic").offsetHeight + 10;

//[...] = ""+width; = ""+height;
document.getElementsByTagName("html")[0].style.width = ""+width;
document.getElementsByTagName("html")[0].style.height = ""+height;

(notice that I get the width and the height of the image and add a small offset to make it visually finer). The fact is that this approach wasn't working either. When opening the extension, one would see it flicker: first a frame of the correct size, and then a resize to the default 10 by 10 pixel frame.

The Solution

The solution was quite simple, but without somebody mentioning it on StackOverflow I would be still stuck. The fact is that the popup of the Chrome Extensions don't resize if there is a hidden element in the HTML tree. That is, basically the whole page I was hiding was preventing me to resize the popup frame. So the solution is:

document.getElementById("hiddenDiv").innerHTML = "";

And that solved every problem. The frame now is the correct size and I can read every two/three days the new xkcd comic.

Friday, February 15, 2013

Newline after every character

I was printing some results for a chart while, after 10 minutes of executing the test, I realised I forgot to put the \n command after the number I wanted to print. I know the numbers I was printing were smaller than 10, so I came up with a useful command to append a new line after each character in a file:

sed 's/\(.\)/\1\n/g' -i filename

And that's it. As easy as that. I'm sorry for the lack of updates but lately I've been very busy.

Wednesday, February 13, 2013

Removing an event listener from an element

Today I was working while I was asked how to remove an event listener from an element which had an
  click event. At first I didn't really remember how to do it, so I Googled a bit and I found this way:

document.getElementById("your_element_id").addEventListener('click', function(){
    //do something
}, false);

As you can see, all the work is done by addEventListener and removeEventListener functions. The first one takes as parameter the event (click in this case) and a function that instruct the event listener on what it has to do with that event. The third argument specifies whether the EventListener being removed was registered as a capturing listener or not (set false by default).

The function removeEventListener takes basically the same input parameters. In the element that has to remove, it specifies which element (click), which function (in this case I use arguments.callee as callee is a property of the arguments object and it can be used to refer to the currently executing function), and the third argument that specifies if it was registered as a capturing listener or not.

Of course there are simpler way to do this. It turned out that who asked me how to remove an event listener from an element was using jQuery. In that case the code looks even simpler:

$('#foo').unbind('click', function() {
   //do something
$('#foo').unbind('click', function() {
   //do something
}, false);

As the previous examples, also in this case there are three parameters. jQuery is smart enough to take all the three possibilities as fine. The first one will remove every listener binded. The second one every function binded to that listener, the third one will remove that particular function binded to that particular event, while the fourth one just specifies the third argument that by default is false.
And that's it. Pretty easy.

Tuesday, February 12, 2013

setInterval/setTimeout and this

I think everybody is familiar with the setInterval and the setTimeout functions. They create a timed event which will occur after a given timeout. setInterval creates an interval: after the given time it will repeat the same action over and over, while the setTimeout function will only execute the action once. Let's see a small example on how they work:

var timed_function = function(){

var timeout = setInterval(timed_function, 1000);

The shown example will execute every second (1000 milliseconds) the function timed_function.

Now, I recently worked with a timed event and I had to access a variable within this. The naive approach was the following:

//previously set variable
this.message = "hello!";

var timed_function = function(){

var timeout = setInterval(timed_function, 1000);

But this printed undefined. I then discovered that when setInterval or setTimeout are executed, the scope of this apparently refers to the window object, thus it's completely loosing the scope of the object in which the function is executed. The idea to solve this problem is to wrap the call inside another object which has to be called to initialise the timeout:

//previously set variable
this.message = "hello!";

var wrapper = function(that) {
    var timed_function = function(){

    var timeout = setInterval(timed_function, 1000);


In this way I basically instantiate the timeout passing through the wrapper function which will keep a reference to this thanks to the input variable of the function. The input variable then will be called inside the timed_function. Hope this helped somebody. I had a couple of issues trying to solve this because I was confused by the scope of the variables. Some folks helped me on StackOverflow luckily, even though some of them where short-sighted and arrogant as always.

Friday, February 8, 2013

Node.JS on Raspberry Pi with OSX, part 3

In this final part I will show how I installed Node.JS on the Raspberry Pi on the Raspbian Wheezy OS.
First of all, open a terminal and execute the following commands:

cd /usr/src
sudo wget

This will change the directory to /usr/src/ and download Node.JS version 0.8.16. I know this is not the last version, but I needed that because my project runs on it and it's not yet time to upgrade to the latest version (even though I'm pretty sure everything is going to be all right).
After the download is completed, extract the tarball:

sudo tar xvzf node-v0.8.16.tar.gz
cd node-v0.8.16

After this you will be in the Node folder, you then should execute the following three commands to build Node.

sudo ./configure
sudo make
sudo make install

This process requires quite some time to complete (especially make). I suggest you don't waste time waiting it to be completed and do something else. Raspberry Pi is not going to break if connected for too much time.
After the commands are finished, Node.JS is ready to be used! You can either run some test with the following commands:

sudo make test

Or just start writing your own web server.

Wednesday, February 6, 2013

Node.JS on Raspberry Pi with OSX, part 2

So after you plug everything (the power supply last), on the screen you should see Raspbian Wheezy booting. Lots of logs in the console and finally what looks like a BIOS screen:

If you don't then you will have the problem I had. In the past hours I struggled a lot because I couldn't see anything on the screen. I tried reinstalling the OS image on the SD (in at least three different ways) and plugging different devices but nothing happened. In the end the problem was indeed the screen (and not the OS as I was supposing). Basically I was using a DELL screen which had a DV input and I was using a HDMI-DV cable. Apparently Raspberry Pi doesn't like that kind of screen inputs. By chance I found a screen with an HDMI input and I could test the Raspberry Pi on it, and it worked!
If you have a different problem I suggest you to try asking on the official forum, there is some helpful people over there.

Anyways, it may also happen that you see something, but not that screen. If that's the case try running the following command:

sudo raspi-config

Then I suggest you set up the keyboard through configure_keyboard and set the timezone through change_timezone. If you happen to have an SD card quite big (more than 4GB at least) then I suggest you also check out expand_rootfs to enable all that space. Also, remember to setup ssh when setting the keyboard by running

sudo setupcon

Then you are good to go! Just hit <Finish> and enjoy your brand new Raspberry Pi. In the next (and hopefully last) post I will show how to install Node.JS and some other useful package.

Tuesday, February 5, 2013

Node.JS on Raspberry Pi with OSX, part 1

During these days I borrowed a Raspberry Pi. I would like to try to set up a Node.js server on it. I decided to write this post to show how I did the whole thing. Just to make things clear, I'm running on OSX.

So first of all once you have your Raspberry Pi, you should mount the OS image on the SD card. Some RPi comes with an SD card with everything setup. If that's the case, you are lucky and can skip this post. Otherwise, just connect the SD card to the machine and mount the OS image on it. To do so you first need to download the image from here. I'm using the  Raspbian Wheezy, which is an optimised version of Debian for RPi. After having downloaded the image, unzip it and download this app. Unzip it and open it. It will ask you to locate the image of the OS first, then you will have to choose the SD card on which you want to mount the image, and finally your password. After it is done, you can remove the SD card and plug it into your RPi. The following image shows how to set up the wirings for your RPi.

Now you have everything ready to start programming on your RPi. In the next post (hopefully) I will show how to set everything up at a software level to start your own Node.js server on it.

Monday, February 4, 2013

Measuring performance of a processor with Node.js

Some time ago I needed to measure the performance of the cores I was using on a machine while a Node.js server was running. I came up with different ideas, including writing a bash script that would measure it the hardcore way. I then realised I could use the os module that comes with Node and use the data it returns to measure it myself.

First of all, take a look at the specification of the method os.cpus() here. It shows how to call it and what is the return value. To measure the performance of the system, it is sufficient to take consecutive snapshots, compute a delta, sum all the data and compute the usage percentage like so:

//include the module
var os = require('os');

//data collected on the previous tick
var previous_cpu_usage;

//[...] collect previous_cpu_usage

//get the cpu data
var cpu_usage = os.cpus();

//iterate through it
for(var index = 0, len = cpu_usage.length; index < len; index++) {
    var cpu = cpu_usage[index], total = 0;
    var prev_cpu = previous_cpu_usage[index];
 //for each type compute delta
 for(type in cpu.times){
     total += cpu.times[type] - prev_cpu.times[type];
 //print it for now
 for(type in cpu.times){
     if(type === "user" || type === "idle")
         console.log("\t", type, Math.round(100 * (cpu.times[type] - prev_cpu.times[type])/ total));

The idea is very simple. Here I'm not showing how to collect data consecutively. This can be done by the means of two timeouts: one that collect one data, then the second collect the second one and then in the callback of the second timeout data gets aggregated. I'm not showing the code exactly as it looks in my solution as it involves message passing through a (big) topology.

In the example shown, I'm printing the data that corresponds to user and idle time, since it's the one most people may be interested in. Anyways, with this code any percentage out of the returned data can be computed.

Sunday, February 3, 2013

Bitwise operations

Some time ago I was talking with a second-year student after a C exam. He was pretty satisfied about the job done, but asked me a question about the ~ operator. At first I didn't quite remember, so I told him that was probably the bitwise not. Later I discovered I was right, so I took the opportunity to refresh my knowledge of the bitwise operations in JavaScript. The purpose of this article is to show which bitwise operations JavaScript offers.

First of all, we can only apply bitwise operations on numeric operands that have integer values. These integer are represented in the operation as a 32-bit integer representation instead of the equivalent floating-point representation. Some of these operations perform Boolean algebra on the single bits, while the others are used to shift bits.

Bitwise NOT ( ~ )

This operator simply flips all the bits of the given integer. In numeric terms, it changes the sign of the number and subtract 1. For example ~0x0000000f will become 0xfffffff0 after the evaluation.

Bitwise AND ( & )

Performs a bitwise Boolean AND operation and set in the result the bit only if it is set in both operands. For example 0xfff666fff & 0x000fff000 will evaluate to 0x000666000.

Bitwise OR ( | )

Likewise the bitwise AND, the bitwise OR behaves in the same way, but a bit is set in the result only if it is set in one or both of the operands.

Bitwise XOR ( ^ )

The bitwise exclusive OR behaves like a normal XOR. The bit in the result is set only if the first operand is true and the second is false, or viceversa. If both are true or both are false, the bit in the result is set to 0.

Shift Right ( >> )

This operator moves all the bits in the first operand to the right by the number of bits specified in the second operand (up to 31). The bits that are shifter off the right are lost, and the filling is given by the initial sign of the first operand (if the operand is positive, it will be zero-filled, otherwise f-filled).

Shift Right with zero Fill ( >>> )

Behaves exactly like the shift right operator, but the filling on the left is always 0, regardless of the sign of the first operand at the beginning of the operation.

Shift Left ( << )

This operator moves all the bits in the first operand to the left by the number of bits specified in the second operand (again, up to 31). The rightmost part of the number is filled with 0s.

These operations are not really used by JavaScript programmers, but it's important to know they exist.

Saturday, February 2, 2013

instanceof & typeof Operators

Sometimes, even if it is considered bad habit, we need to check the class of an object in JavaScript. Classes may be defined by the programmer, but I will cover this topic later, for now let's just focus on classes defined by the interpreter. There are many, for example:
  • Object
  • String
  • Number
  • Date
  • Function
  • ...
To check if an object is an instance of a specific class, JavaScript offers the instanceof method, which may be used in the following way:

var mynmbr = new Number(1);
var mystrng = new String("disasterjs!");
var myfunc = function(){
    console.log("hello world!");
mynmbr instanceof Number //returns true
mynmbr instanceof Object //returns true
mystrng instanceof String //returns true
mystrng instanceof Object //returns true
mynmbr instanceof String //returns false
myfunc instanceof Function //returns true

Be careful when using instanceof about what you are checking. For example numbers and strings not wrapped in the constructor are treated as primitive value, thus a call for instanceof in that case would produce an unexpected result:

var mystrng = "disasterjs!";
var mynmbr = 1;

mynmbr instanceof Number //returns false
mynmbr instanceof Object //returns false
mystrng instanceof String //returns false
mystrng instanceof Object //returns false

What about the typeof operator? This operator returns the class as a string. The following table summarises its possible return values:

Host object (provided by the JS environment)Implementation-dependent
E4X XML object"xml"
Any other object"object"

Programmer may be careful here when checking a Null object as the result of a typeof will return "object". Here are a couple of examples on how to use typeof.

typeof 1 === "number" //true
typeof Number(1) === "number" //true
typeof 1 === "object" //false
typeof undefined_variable === "undefined" //true
typeof true === "boolean" //true

And that's it.

Friday, February 1, 2013

JavaScript, passing-by-reference and slice()

Some time ago I had a problem with arrays and assignments. I was working with a distributed system in JavaScript; the root process sent some "work" to do to some subprocesses (workers) and those returned an array as result, which was modified inside the callback of the dispatching function.

The issue was that when I modified the array, I then assigned it to a variable and passed it to a function. The problem was that while I was working in the callback with this array, it got modified somehow. I later understood that the problem was due to the fact that JavaScript passes values by reference. That is basically the array I assigned to the variable before passing it to the function was simply saving it's reference on the variable, but not really duplicating it.

The following example will show the problem.

var arr = [1, 2, 3]

var arr2 = arr;

arr2[0] = 0;


//prints 0

As shown, this prints zero. To solve the issue I used a custom procedure to copy every single value of the array into a new one. This is pretty painful and later I discovered I could simply use the slice() function. The following snippet based on the previous example will show the black magic:

var arr = [1, 2, 3]

var arr2 = arr.slice();

arr2[0] = 0;


//prints 1

Basically that's it. It's a very naive problem, and a skilled programmer may quickly find the issue, but knowing the existence of the slice() function may save quite some time.

How to syntax-highlight code on Blogger

One of the first issues I had with this blog was how to syntax-highlight code snippets. All I could find on the internet was garbage and old tutorials with the previous Blogger interface. Then I finally found this blog that proposed a working solution. Actually, it's the only solution that worked, so I'm suggesting it here, but I'm not entirely satisfied with the CSS that comes with it, so I might slightly modify it sooner or later.
If somebody knows a better solution, let me know in the comments, I'll be glad to take a look at it.