Catching up!

A lot has changed since the last blog post (more than three years). I was happily running a successful business around Videocache till Google decided to push HTTPS really hard and enforced SSL even for video content. That rendered Videocache completely useless as YouTube video caching was the unique selling point. Though people are still using it for other websites (whatever supported and not HTTPS yet), I personally didn’t find it good enough for selling. To add to the trouble, Mozilla and friends announced that there will be free certs for everyone. Now, that took away whatever motivation was left to keep working on Videocache. I decided to open source Videocache and the source is now available on GitHub. If you have better ideas or you are looking forward to make things work by forging certs etc, fork it and give it a shot.

After all that mess with Videocache, I am left with a contract job which is not working out well. So, the learning hat is back and I am trying to catch up with the tech world. I didn’t really want to get into the whole JavaScript framework mess because there is no clear winner and there are too many of them but it looks rather unavoidable or Unflushable (if you remember Jeff from Coupling). I have been trying to make simple apps with Angularjs. There is little documentation and you have to dig really deep at times but still you can make things work if you persist. Once you spend some quality time with it, you may actually start liking Angularjs. So, I did give it some time and implemented angular fronted for Pixomatix-Api (another learning project I am working on) and I think I sort of like it now.

In 2015, if you are web developer, you must know how APIs work and you should be able to consume them. So, to learn to expose APIs and version them properly, I fired a small project Pixomatix. Being a Rails developer, you really get obsessed with it and try to implement everything using Rails. Even when you want an API with 2-3 endpoints, you tend to make the horrible mistake of doing it in Rails. This kept bugging me and a few weeks later, I decided to freshen up my Sinatra memories. But working with Sinatra is not all that easy especially if you are used to all the niceties of Rails. Dug up my attempt of implementing Videocache in Ruby, and extracted few tasks and configurations I had automated long time ago. Ended up working a lot more on it and packages into a template app with almost all the essential stuff. Though I need to document it a little more, the app has got everything needed to expose a versioned API via Sinatra.

On the other hand, I tried to use devise gem to authentication for Pixomatix. It was all good for integration with standard web apps and APIs but it sort of failed me when I tried to make the API versioned. Devise turned out to be black-hole when I tried to dig deeper to make things work. I tried a few other gems which supported token authentication but they were also no good for versioning. Generally, you may not need to version the authentication part of your API, but what if you do! Since, this was just a learning exercise, I was hell bent on implementing this. So, I just reinvented the wheel and coded basic authentication (including token authentication) for the API.

That’s it for this post. I am looking forward to post regularly on the new stuff I learn.


Javascript To Remove All Facebook Apps From Your Account

This is a guest post by Gaganpreet Arora AKA bitgeek.

My Facebook app list was cluttered with useless apps accumulated over the time I had joined Facebook. The worst thing is that Facebook does not allow you to remove multiple apps in one go. You have to select one, confirm on “Are you sure?” first and then click on “Okay” when it is deleted. Very cumbersome. So I wrote a Javascript hack to remove all of the apps in one go.

Warning: This will actually delete all the apps

Step 1

Go to:

Step 2

Select and copy the complete code below and paste it into the address bar.

function postwith (to,p)
  var myForm = document.createElement("form");
  myForm.method="post" ; myForm.action = to ;
  for (var k in p)
    var myInput = document.createElement("input") ;
    myInput.setAttribute("name", k) ;
    myInput.setAttribute("value", p[k]);
    myForm.appendChild(myInput) ;
  document.body.appendChild(myForm) ;
  myForm.submit() ;
  document.body.removeChild(myForm) ;
function getElementsByClassName(classname, node)
  if (!node)
    node = document.getElementsByTagName('body')[0];
  var a = [], re = new RegExp('\\b' + classname + '\\b');
  els = node.getElementsByTagName('*');
  for (var i = 0, j = els.length; i < j; i++)
    if ( re.test(els[i].className) )
    { a.push(els[i]); }
  return a;
var arr = [];
arr['remove'] = 1;
arr['post_form_id'] = document.getElementById('post_form_id').value;
arr['__a'] = 1;
for (i=0; i
<h3><span style="text-decoration: underline;"><strong>Step 3</strong></span></h3>
Press Enter.
The browser may get stuck for a few moments if the list is large, so if it does, just wait a bit and it will respond back.
<h3><span style="text-decoration: underline;"><strong>References</strong></span></h3>
Function postwith was taken from <a href="">here</a>, and getElementsByClassName was taken from <a href="">here</a>.

News : Memepress (Yahoo! Meme) WordPress Plugin Version 0.3 is available

Memepress is a wordpress plugin which provides wordpress widgets for displaying public posts from Yahoo! Meme. Memepress is SEO ready and provides options to noindex/nofollow your Meme posts. New version fixes a few bugs and provides enhanced style control for the widget.

You can download memepress version 0.3 from wordpress plugin page.

For details on usage/installation instructions, please check plugin homepage.


How To: Install PHP 5.3 on CentOS 5.1 or RHEL 5.1

PHP 5.2 and PHP 5.3 has tons of functionalities which are not provided by PHP 5.1 and PHP 5.1 is the most recent version of PHP available in CentOS 5.1 and RHEL 5.1 via updates. Also, its difficult for people to upgrade OS(s) on production servers frequently and especially if the servers are running absolutely fine, it doesn’t really make sense to upgrade just for the heck of it. One more reason for no upgrades is the OS choices provided by many hosting companies. But a lot of CMS(s) and softwares are now recommending at least PHP 5.2 for an extended feature set.

I tried a lot of repos for CentOS including RPM Fusion, RPM Forge, EPEL etc., but none of them could provide even PHP 5.2. In some corner of my mind, I had a faint idea of Remi Collet publishing updates about Remi repository. Did a quick google search and found out that PHP 5.3 is actually available in the repository ๐Ÿ˜€

So, if you are really in need of PHP5.3 on your CentOS/RHEL 5.1 or lesser, head over to Repository Configuration Page and follow the instructions.

Note: The repository is not enabled on install (enabled=0). You need to enable it in /etc/yum.repos.d/remi.repo.

Once installed and enabled, you can update PHP, MySQL from remi repository using the following command

[root@fedora ~]$ yum update php php-* mysql-*

Warning: Drupal is still not totally compatible with PHP5.3 and if you run Drupal sites, be sure to investigate things before upgrading.


Memepress Ready to be Translated (Supports Turkish now)

Thanks for an awesome response to the Memepress (Yahoo! Meme) WordPress plugin. A lot of people have blogged about it and contacted me via email for enhancements. New version also fixes the bug regarding image size in posts. I have also updated the plugin to support internationalization(i18n). Now you can translate about 20 strings in your local language and you’ll have Memepress in your own language.

Turkish Translation

Huge thanks to Yalรงฤฑn Erdemir for translating the plugin to Turkish :)

How To Translate

If you are interested in translating Memepress plugin in your language, then download the memepress.pot (POT file) and open it in your favourite text editor. Now for every string against msgid should be translated to your language and populated in quotes against msgstr. See the example below.

Before translation

#: memepress.php:37
msgid "Title"
msgstr ""

After translation

#: memepress.php:37
msgid "Title"

Once you are done with translating all the strings, save the file and mail me at kulbirsaini25 [AT] .


How To: Configure Dual Display with ATI Radeon (fglrx)

As promised in my last post (News: ATI Catalyst Display Drivers 9.9 Released), I am back with a post on configuring dual display with ATI Radeon HD Graphics Card and proprietary catalyst (fglrx) drivers from ATI.

Hardware Used

Graphics Card: ATI Radeon HD 3200 (256MB, onboard)
Monitor 0: ViewSonic VG1930WM 1440×900 (19″ LCD, Connected via DVI port)
Monitor 1: Samsung SyncMaster 793S 1280×1024 (17″ CRT, Connected via VGA port)

Types of Dual Display

  • Mirror: Both screens have same content, identical refresh rate and resolution.
  • Clone: Both screens have same content but refresh rates and resolutions can be different.
  • Horizontal: Both screens can have different content, refresh rates and resolution. Screen 1 is left or right of Screen2.
  • Vertical: Same as horizontal. The only difference is that Screen1 is above or below Screen2.

In this post, we are interested in Horizontal setup with xinerama on. This way we can have two desktops allowing full screen modes on both of them and allowing us to drag and drop windows from one screen to the other.

Install ATI Drivers

If you don’t have ATI drivers installed already, follow this How To: Install ATI Catalyst (fglrx) Drivers on Fedora 11 (works for any version of ATI Catalyst drivers).

Generate xorg.conf file

If you don’t see the xorg.conf file at /etc/X11/xorg.conf, then you need to generate it to proceed to next step. Use the following command as root to generate one

[root@fedora ~]$ Xorg -configure

This command will generate the default xorg.conf file at /root/ Copy it to /etc/X11/xorg.conf.

[root@fedora ~]$ cp /root/ /etc/X11/xorg.conf

Backup xorg.conf file

Backup your original xorg.conf file so that you can restore it in case the configuration doesn’t work the way you expected.

[root@fedora ~]$ cp /etc/X11/xorg.conf /etc/X11/xorg.conf.backup

Generate Configuration for Dual Display

Now we are ready to generate the configuration for dual display. Use the following command with appropriate arguments (in accordance with your hardware configuration)

# --screen-layout will place second screen on left of your first screen. Other possible values are right,above,below.
# --xinerama=on option enables you to have two different desktops and one of them being passive.
# You can drag and drop windows from one desktop to the other. Task bars appear only on one of the desktops.
[root@fedora ~]$ aticonfig --initial=dual-head --screen-layout=left --xinerama=on
# --resolution=_screen_number_,widthxheight
[root@fedora ~]$ aticonfig --resolution=0,1440x900 --resolution=1,1280x1024
# Set horizontal sync and vertical refresh rates for both monitors.
[root@fedora ~]$ aticonfig --hsync=0,30-60 --hsync=1,30-60 --vrefresh=0,30-60 --vrefresh=1,30-60

You can download my xorg.conf file via this link.

Reboot or Logout and Login Again

If you just setup your ATI drivers and configured the dual display, you need to reboot so that fglrx module can be loaded properly. If you rebooted after setting up the drivers, just logout and login again to checkout your dual display :) If everything works fine, say thanks to me and if not blame ATI ๐Ÿ˜›

Adjust DPI for Normal Font Size

I faced a problem with my font sizes being too big while using xinerama. It was easy to fix by adjusting DPI. Go to System -> Preferences -> Appearance. Go to Fonts tab. Click Details located near the bottom right corner. On that window, try descreasing the “Dots Per Inch” value. Mine worked fine with 85 DPI.

Below is an image of my dual display setup. Click to enlarge.

Dual Display Configuration ViewSonic Samsung Using ATI Radeon Catalyst (fglrx)


News: Memepress (Yahoo! Meme) WordPress Plugin is out!

Yahoo! Meme is a new Twitter like service in testing by Yahoo!. But its much better than Twitter. It has got facility to post text, photos, video and music. The best thing about Meme is the user interface. I am just loving it. Such a clean interface :) I liked twitter’s web interface very much because it was not cluttered like facebook and others. But meme’s interface is much better than twitter’s interface.

As I felt hooked to it, I thought it’ll be a good idea to have wordpress widget to display latest posts from Yahoo! Meme. I took Twitter for WordPress plugin’s source code and started modifying it to work for Yahoo! Meme and in a few hours, I managed to get a full fledged plugin for Yahoo! Meme and a few extra features and minor bug fixes as well :)

I have named it Memepress. For more details and latest updates about the plugin, please visit plugin homepage on wordpress. The plugin is really easy to setup and doesn’t have any dependencies and doesn’t even need to authenticate with Yahoo! Meme.


Note: Yahoo! Meme is still invite only. If you need an account, send me an email using this contact form.


How To: Install and Use Twython (Python Wrapper for Twitter API)

As promised in my previous post, here is a brief howto on getting started with twython. The main advantage of Twython over several other python (or any other language) wrappers for Twitter API is that it works even when you are behind your organizations proxy.

Download Twython

You can download latest version of twython from twython page on github. You can either clone using git (if you have git installed) or can click the download button.

Install Twython

Once you are done with extracting the downloaded tar file. Change directory to twython and run these command as root.

[root@fedora ~]$ git clone git://
[root@fedora ~]$ cd twython/dist
[root@fedora dist]$ tar -xvzf twython-0.8.tar.gz
[root@fedora dist]$ cd twython-0.8/
[root@fedora dist]$ python build
[root@fedora dist]$ python install

Use Twython from Python Interpreter

Below is a direct copy paste lines from my interpreter. See how things are working (learning by doing).

[saini@bordeaux ~]$ python
Python 2.6 (r26:66714, Mar 17 2009, 11:44:21) 
[GCC 4.4.0 20090313 (Red Hat 4.4.0-0.26)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> # first of all, import twython module
>>> import twython
>>> # Authenticate your twitter account with your twitter username
... # and password with twitter.setup method.
>>> client = twython.setup('Basic','myusername','mypassword')
>>> client.authenticated
>>> #Lets update our current status on twitter with some cool message.
>>> client.updateStatus('Testing #twython. The coolest #TwitterAPI :)')
>>> # Now go and check your current status on twitter. Surprised!!!
>>> # Get your or anyone's followers
>>> client.getFollowersIDs(screen_name='gofedora')
>>> # Output truncated.
>>> # Get help for any function.
>>> print client.createFriendship.__doc__
createFriendship(id = None, user_id = None, screen_name = None, follow = "false")
	Allows the authenticating users to follow the user specified in the ID parameter.
	Returns the befriended user in the requested format when successful. Returns a
	string describing the failure condition when unsuccessful. If you are already
	friends with the user an HTTP 403 will be returned.
		** Note: One of the following is required. (id, user_id, screen_name)
		id - Required. The ID or screen name of the user to befriend.
		user_id - Required. Specfies the ID of the user to befriend. Helpful for disambiguating when a valid user ID is also a valid screen name. 
		screen_name - Required. Specfies the screen name of the user to befriend. Helpful for disambiguating when a valid screen name is also a user ID. 
		follow - Optional. Enable notifications for the target user in addition to becoming friends. 

So now you are ready to do wonders with twython. Write your own code and blog/brag about it :)


I Love Twitter, I Love Python = I Love Twython

I was really late to join Twitter. As soon as I joined Twitter, I started exploring various aspects of Twitter and hence the Twitter API. While searching for a wrapper for Twitter API in python, I came across python-twitter which was good and tweepy which was ok and I finally found twython. Twython is awesome and more importantly upto date with Twitter API.

While I was using twython, I found one fact very irritating. It didn’t have docstrings. And I had to go to Twitter API Wiki for description of every single function :( I talked to Ryan (Twython developer, who works all day long at his day job with and develops twython when everybody sleeps ๐Ÿ˜› ) about the same and offered to write docstrings for twython.

I spent a few hours and sent a patch with docstrings for all the functions. Now twython has complete documentation. Also Ryan is trying his best to implement wrappers for the newly introduced functions in Twitter API. So, currently Twython is THE best wrapper available in python for Twitter API.

PS : Next post will be “How to use twython”.


How To: Recover Deleted Files in Linux Using Photorec

Frequently I get messages like “I have accidentally deleted my project. Is there any way to recover it in Linux?” or “I worked really hard on that program and managed to deleted it right before submission :( How do I recover?”. Well I didn’t really have any idea on “How to actually recover deleted files in Linux”.

A few days back I was designing the new logo for Videocache in Inkscape. After finishing the design, I saved the svg file carefully. And deleted other images which I embedded in the logo. A few moments later I realized that I just screwed up myself by deleting all those files as the svg file is now good for nothing.

I searched a lot on recovering files in Linux but in vain. Then a friend (bitgeek) told me about Photorec. I managed to recover all the files using photorec. I thought it would be a good idea to let others know and spread a good word about Photorec. Below is a step by step howto on using Photorec to recover your files.

What is Photorec?

From Photorec website,

PhotoRec is file data recovery software designed to recover lost files including video, documents and archives from Hard Disks and CDRom and lost pictures (thus, its ‘Photo Recovery’ name) from digital camera memory. PhotoRec ignores the filesystem and goes after the underlying data, so it will still work even if your media’s filesystem has been severely damaged or re-formatted.

For This HowTo

Lets say I had a file download_arrow.png in /home/saini/Desktop which I have removed accidentally. Login as root and create a directory recover which will be used to store all the recovered files.

Install testdisk/PhotoRec

Photorec comes as a part of testdisk package in Fedora (I hope its same for other distributions as well). Use yum to install testdisk.

[root@fedora-tips ~]$ yum install testdisk

Launch Photorec

Once you are done with installation. Open a terminal and launch photorec (as root).

[root@fedora-tips recover]$ photorec

Select Hard Disk

If you have more than one hard disk in your system, select the one from which you have deleted the file(s).

PhotoRec Hard Disk Selection

Select Partition Type

If your hard disk has Linux partitions, then select [Intel].

PhotoRec Partition Selection

Select Filetype Option

Move to [File Opt] and press enter. Here you can disable all file types by pressing ‘s’ . Use space to toggle the check button. Now since we removed a png file, we are going to check only png file type.

PhotoRec Filetype Selection

Select Options

Photorec also has a list of different options. Under normal circumstances you don’t need to modify them

PhotoRec Option Selection

Select Partition

Move the selector to the partition from which you have removed the file. Then press enter on search.

PhotoRec Partition Selection

Select Filesystem Type

If you are using Linux, its going to be ext2/ext3/ext4. So the default selection is file.

PhotoRec Partition Selection

Select Space for Analysis

Select free if you didn’t write to that partition after removing the particular file otherwise select whole.

PhotoRec Space Selection

Select a Directory to Recover Files

Now select the path where the recovered files will be stored. Then press ‘Y’.

PhotoRec Recovery Directory Selection

Recovery Progress

Photorec will show how many files it has recovered.

PhotoRec Recovery Progress

All recovered file will be stored in the directory selected above. Open them in a file browser and you’ll get the removed file there. I hope this howto will help you recovering files you accidentally delete :)