Just for fun :)

This is something I did last year and had completely forgotten about it till today. Was bored and sitting in front of my laptop wondering what to do when I remembered this.

On a similar incidence last year, I ended up writing a simple script to play rock paper scissor on gtalk. Yes I know this sounds completely hilarious, but given the circumstance and the boredom, it was a fun thing to do as a programmer.

It was a not-so-interesting day at office and was chatting with my friend, talking about our usual non-sense  until we came to the fact "...what if one of us wasn't there to talk to when the other got bored ?"...well, I said, let's write a script that replies back to the other as us. since we know each others pattern very well, it should be possible... But wait, what if, both of us setup our chat bots and one of us initiates it. what if, I initiate the first chat msg, then your bot will reply as you, then my bot will receive it and reply as me, and then your bot and then mine....and so on...a non-ending loop!, we asked each other.

As crazy as it seemed, I was bored and this thought invoked the curious geek in me. I wanted to do it and check it out, if not for anything, atleast for the fun of it. So, off to it then.

I use linux with awesome wm  and mcabber as the chat client. Both awesome and mcabber provides a scripting interface to control them and one could do some amazing things with it. Those who've seen & been with me already know my gtalk and other notification integrations [1] with awesome. So all I had to do now was to write up another script for this. But then, a script that mimics me!!...what was I thinking of...I'd be competing in the turing's test then. Soon, with my senses back to normal, it was clear that I just needed to script that would reply something in a predefined pattern. For all that mattered, even a line of script that sends a char back and forth would do the job but then that would take out all the fun in it. So I turned to the most common game that kept me occupied at classes during the school days..."rock, paper, scissors". After a brief nostalgic moment, started implementing it. Simple logic, few lines of "if" and "else" statements with some pattern matching and it was done.

All that one had to do was ping me a line of chars, like, "rps= rock;". (rps=rock, paper, scissor :). With those chars, my mcabber script would parse, analyze it and reply back with msgs. Here's how it looked like.

It soon turned out to be a fun time when my friend joined in to play along. She put it in an infinite loop and soon my gtalk was flooded with these txts. Mcabber handled it, but gtalk and the hangouts app crashed across all my devices. On the phone, tab, chrome & firefox on both my laptop and my workstation. Guess we had hit the limits. Next I wanted to make it "rpsls"....Yes. Sheldon's rock, paper, scissor, lizard, spock!!" \\\\//,

And so, we got lost in couple of this and that and digressed away from our original intention.

Finally after many months, another boring night with nothing much to do. Sitting in front of my laptop, my friend's pic on gtalk reminded me about this rps that I had done. I must admit that it didnt seem as much exciting now as it had back then. Nevertheless, I gave it a shot to complete what I had started long ago. Quickly setup it up with my gmail and another temp account. With each mcabber client side-by-side, I initiated the first cmd "rps= rock;", and voila...was working as expected.

It was nice to see the two clients play with each other. Gave that whole satisfaction and happiness of coding for the fun of it. Also brought a whole range of fun ideas that I could use this for (thinking of remotely controlling my media player through gtalk..etc and what not). Guess it would have been even more fun if my friend was there online, to let the bot run on her system and mine here. To see our bots play and compete with each other while we spoke of our usual non-sense :)

Wondering what more I could do with this. Just a few days back it was officially declared that a super computer passed the turnig test. Not a bad thought eh...may be the next time I'm bored, I can work on building a client that imitates me. Would be good fun.

[originally posted in july,2012 on my blogspot]

ActiveMQ is one of the popular open source messaging and integration patterns server. It supports many cross language clients and protocols. One of it's feature is the auto-mapping between multiple protocols. 

But one might encounter issues when using multiple clients with stomp and mqtt protocols with ActiveMQ when dealing with path separators for topics.

The "DestinationPathDescriptorPlugin", is an activemq plugin that can be used to resolve the issues.

What is this plugin for ?

Consider this scenario. Two clients, one using stomp and other using mqtt protocol, need to communicate through a common topic named "foo/bar" using ActiveMQ as the message broker. 

The stomp clients uses "/topic/foo/bar" for the topic name and the mqtt client uses "foo/bar". This is well and good according to the standards of each protocol. ActiveMQ on the server side, converts the path separator of mqtt from "/" to ".", for making it compatible with JMS as JMS uses "." for path separator, but doesn't convert the path separator of stomp and keeps it as "/" itself.

The end result is that there are two different topics created, "/topic/foo/bar" and "foo.bar", on the server side and because of this, the two clients will not be able to communicate with each other. But what is required would be the creation of a single topic and for activemq to automatically map both the stomp and mqtt to the same topic so the clients can communicate successfully.

Although, this seems a bit contradictory to the activemq's feature, one might face this issue in reality and scarcely fathom what might be the problem.

One simple solution would be to use the "DestinationPathDescriptorPlugin" provided by activemq, which allows clients to use custom path separator. By using this, whatever be the path separator used by the clients, on the server side the path separator that would be used is the one that is defined by the plugin. By default, the plugin provides "/" as the path separator.

So, in our scenario, whether the mqtt client is using "/" or "." as the path separator and the stomp client is using "/" as the path separator, on the server side the topic created would use "/" as the path separator and hence a single topic would be created and now both the clients can successfully communicate with each other.

How to get the plugin

The plugin comes bundled by default with the activemq package (at least with versions >=5.6).

To use the plugin, simply add the following into your activemq.xml conf file. (This by default uses "/" as the destination path separator)

<broker> ....... <plugins> ........ <destinationPathSeparatorPlugin/> </plugins> </broker>

Extra info

According to the official documentation, this should be added last if you want it to be working properly with other plugins.

[originally posted in dec, 2012 on my blogspot]

Recently I got a new android phone for myself unexpectedly. I already had one earlier and now i was wondering what i would with the two phones. I just didn't want to give away the older android phone so i began to wonder what use i can put the old phone to. Many ideas came to mind but the one i found most interesting was to somehow make it work like ip camera and see whats happening on my laptop and other devices connected on my network.

So me being an android developer, all excited with the idea, hastily started off with building an app to steam the video from the phone. After a few lines of code i came to senses and thought may be i should look out if any such apps were already available. And not so surprisingly there were. Among the few i found the best one was "IP Webcam". It did exactly what i wanted, stream the video feed form the phones camera over a network. It worked very well and satisfied the requirements i had initially but i was missing the whole joy of me getting to code and my code doing something.

After a few tries, the whole setup became boring. There was just one frame on my laptop with the video feed from the camera and nothing much special about it.So I thought I'll put more liveliness to it by sending notifications when something happens...in this case, i wanted notification when some one came near the phone so that I'd now know that there is someone/something near my phone and now the video feed would be good to see. A simple notification like "There is something/someone in front of the phone. Want to see it- yes/no?".

This time i knew i had to code and wouldn't find something this customized on-line. My plan was to use the proximity sensor on the phone to detect some presence and use tcp/ip sockets to send a notification to my system. This would be my client and a small sever code on the laptop would be listening and when it gets this msg from phone would pop up a notification with "yes / no" and on

<<<paaa"yes" would open up the video stream. Something similar to this.

So, the first piece of the puzzle is to setup a simple server-client communication between the phone and the laptop. The few lines of simple java code does the job

client:

public class CameraClient { public static void main (String args[]) {// arguments supply message and hostname of destination Socket s = null; try{ int serverPort = 6880; s = new Socket("localhost", serverPort); DataOutputStream output = new DataOutputStream( s.getOutputStream()); //send some data System.out.println("sending data"); output.writeInt(6); } catch (UnknownHostException e){ System.out.println("Sock:"+e.getMessage());} catch (EOFException e){ System.out.println("EOF:"+e.getMessage()); } catch (IOException e){ System.out.println("IO:"+e.getMessage());} finally { if(s!=null) try {s.close(); } catch (IOException e) {/*close failed*/} } } }

Server:

public class CameraServer { public static boolean notify = false; public static void main (String args[]) { try{ int serverPort = 6880; ServerSocket listenSocket = new ServerSocket(serverPort); System.out.println("server start listening... ... ..."); while(true) { Socket clientSocket = listenSocket.accept(); Connection c = new Connection(clientSocket); } } catch(IOException e) { System.out.println("Listen :"+e.getMessage());} } } class Connection extends Thread { DataInputStream input; Socket clientSocket; String clientIP=""; public Connection (Socket aClientSocket) { try { clientSocket = aClientSocket; input = new DataInputStream( clientSocket.getInputStream()); InetAddress i = clientSocket.getInetAddress(); if(i !=null){ clientIP = i.toString().substring(1); System.out.println("connected to " + clientIP); } this.start(); } catch(IOException e) { System.out.println("Connection:"+e.getMessage()); } } public void run() { try { // an echo server //Step 1 read length int nb = input.readInt(); System.out.println("Got notification "+ nb); //Step 2 read byte if(CameraServer.notify == true){ System.out.println("Notification in progress..."); return; } CameraServer.notify = true; JDialog.setDefaultLookAndFeelDecorated(true); int response = JOptionPane.showConfirmDialog(null, "There seems to be some activity in sector1.\nDo you want to see video feed?", "Confirm", JOptionPane.YES_NO_OPTION, JOptionPane.QUESTION_MESSAGE); if (response == JOptionPane.NO_OPTION) { System.out.println("No button clicked"); CameraServer.notify = false; } else if (response == JOptionPane.YES_OPTION) { System.out.println("Yes button clicked"); CameraServer.notify = false; Runtime run = Runtime.getRuntime(); Process pr = run.exec("/usr/bin/cvlc \"http://"+clientIP+":8080/videofeed\""); } else if (response == JOptionPane.CLOSED_OPTION) { System.out.println("JOptionPane closed"); CameraServer.notify = false; } } catch(EOFException e) { System.out.println("EOF:"+e.getMessage()); } catch(IOException e) { System.out.println("IO:"+e.getMessage());} finally { try { clientSocket.close(); } catch (IOException e){/*close failed*/} } } }

The server listens on a particular port and for every connection, spawns a thread, give a notification and if "yes" is clicked opens up the video feed in vlc.

The next part was to interface with the proximity sensor on the phone. That's pretty simple and lots of examples on-line like the following code snippet from http://android-er.blogspot.in/2011/09/monitor-proximity-sensor.html

SensorManager mySensorManager; Sensor myProximitySensor; @Override public void onCreate(Bundle savedInstanceState) { super.onCreate(savedInstanceState); setContentView(R.layout.main); ProximitySensor = (TextView)findViewById(R.id.proximitySensor); ProximityMax = (TextView)findViewById(R.id.proximityMax); ProximityReading = (TextView)findViewById(R.id.proximityReading); mySensorManager = (SensorManager)getSystemService( Context.SENSOR_SERVICE); myProximitySensor = mySensorManager.getDefaultSensor( Sensor.TYPE_PROXIMITY); if (myProximitySensor == null){ ProximitySensor.setText("No Proximity Sensor!"); }else{ ProximitySensor.setText(myProximitySensor.getName()); ProximityMax.setText("Maximum Range: " + String.valueOf(myProximitySensor.getMaximumRange())); mySensorManager.registerListener(proximitySensorEventListener, myProximitySensor, SensorManager.SENSOR_DELAY_NORMAL); } } SensorEventListener proximitySensorEventListener = new SensorEventListener(){ @Override public void onAccuracyChanged(Sensor sensor, int accuracy) { // TODO Auto-generated method stub } @Override public void onSensorChanged(SensorEvent event) { if(event.sensor.getType()==Sensor.TYPE_PROXIMITY){ ProximityReading.setText("Proximity Sensor Reading:" + String.valueOf(event.values[0])); } } }; }

Ha, with that done the only thing left to do was to glue it up all together and put some android specific code and configurations to make the sensor app run in background and send notification on sensor events. Works beautifully now. Although the proximity sensor on the phone ain't that great. May be in time i shall hook up a proper sensor and interface it with the phone. But never mind that, for me my old phone is again useful in a new cool way and I had my cup of joy for a Sunday.

[originally posted in oct, 2012 on my blogspot.]

There are times when i get totally frustrated due to some networks limiting my download capabilities. Mostly my college network!!. Had to find an alternative solution but the fact that i'm almost always on cmdline using wget or curl kept me away from using any of those fancy UI based download managers which claim to accelerate downloads. Well a bit of quick googling pointed me to "axel"... a simple yet wonderful cmdline tool that does exactly what it is meant to in a most elegant way. For those curious few, here is its residence http://axel.alioth.debian.org/ 

For those die hard fans of wget & curl, axel would be an immediate replacement. Its as simple as

$axel -n 10 'http://www.kernel.org/pub/linux/kernel/v3.0/linux-3.6.3.tar.bz2'

this starts 10 simultaneous connections to get the files from the server.

While this would suffice most of the time, another pain is during system updates when the package manager defaults to using wget. So here are a few well know tricks to make the package manager use axel rather than wget.

apt-get users:

Download this nice little script by Matt parnell from here.

and then

#mv apt-fast.sh /usr/bin/apt-fasht && chmod +x /usr/bin/apt-fasht

Give it any fancy name you want and start using it instead of apt-get.

pacman users:

use the following Xfercommand in your /etc/pacman.conf

XferCommand = /usr/bin/axel -n 2 -v -a -o %o %u

emerge users:

add the following line to your /etc/make.conf

FETCHCOMMAND="/usr/bin/axel -a -o /\${DISTDIR}/\${FILE} \${URI}" RESUMECOMMAND="/usr/bin/axel -a -o /\${DISTDIR}/\${FILE} \${URI}"

here are some samples...

[original posted on may,2011 in my blogspot]

 Recently, I was thinking about the way we write c programs. There are certain set of statements and a few other things that we are bound to use again and again. Like the main routine, the headers, the comments...to mention a few. It would be nice if the editor did these redundant tasks. We would code lot faster. Such features are normally provided in the IDEs, but being a linux user I use vim to code c programs. Vim has so many features, i was sure that it would also have this...and I was darn right about it :)

'c.vim' is a small plugin for vim that converts it into a c/c++ IDE. Adding this to vim considerably increased my coding speed. It provides insertion of complete statements, code snippets, comments and a few more nice features.

Its fairly easy to change the template for the code snippets and the comments. You can also set your personal details like your name to be displayed, your contact etc. You can follow the documentation that is provided with the plugin to customize it to your needs. Once that is done, just take a look at the key combination to perform the actions.

Now open a new c file to code as usual, and you'l automatically be greeted with your information (the personal details which you've set in the template file) and also an empty main routine. The cursors is also placed in the appropriate positions so that you can continue to code.

All the key combinations begins with '\'. for example, '\cfr' creates a frame comment. There are many more available. Make sure to read the 'c-hotkeys' file that is provided with the documentation. Download the c.vim here

...being a linux user I've tried out various linux distros, trying to find the one that suits me the best.

Each of them are wonderful in their own way. Many of the distros are made for a specific set of users. Some aim at security, some at being small and off course some to replace windows :)

 But none of them seem to satisfy me. Some distros provide wonderful UI and customization but hog a lot of memory. Few of them are very fast and small but does not provide a good GUI. Most of the time, there are so many unwanted features that I hardly use.

So I decided that I'l build my own based on my requirements. In the beginning I thought of using 'Arch' as it comes with a plane and simple base system upon which I can add my required modules. Moreover since 'Arch' is based on rolling release, the fact I would not have to frequently upgrade to newer stable versions was comforting.

Then I came across this project call 'linux from scratch'(LFS). I'm glad I found that project. LFS provides step-by-step instruction to build a custom linux from source code. I couldn't ask for more. This is the next best thing to writing my own OS.

LFS lets me create a basic system. similarly there are other xLFS, that lets you to continue on the base system to create more advanced and customized systems .

lost your grub after installing windows? here is a simple way of getting it back.

[note: assuming windows being installed on top of ubuntu(using grub2) ]

1: boot from a live cd and open a terminal

2: sudo fdisk -l

this will show the partitioin table. Here is my partition table

Device Boot Start End Blocks Id System /dev/sda1 63 771119 385528+ de Dell Utility /dev/sda2 * 772096 105629695 52428800 83 Linux /dev/sda4 105631722 625141759 259755019 f W95 Ext’d (LBA) /dev/sda5 105631744 315355949 104862103 7 HPFS/NTFS /dev/sda6 315356013 561134384 122889186 7 HPFS/NTFS /dev/sda7 621142016 625141759 1999872 82 Linux swap / Solaris /dev/sda8 561134448 621137159 30001356 83 Linux

3: mount the linux partition on which you want install the grub (sda8 in my case)

$sudo mount /dev/sda8 /mnt/ $sudo mount --bind /dev /mnt/dev $sudo mount --bind /proc /mnt/proc

4: Now chroot into this environment

sudo chroot /mnt

[optional]if you want to change any default setting like timeout optiions... then look at /etc/default/grub

5: #grub-install /dev/sda

[ in case of error do #grub-install --recheck /dev/sda ]

6: now unmount and reboot the system.

I am a cmd line user, so i spend most of my time on the terminal and its really painful to use the gui for some of the common things. Power management is one such case.

Just to put the system to sleep i've to go to the menu and then select, suspend. Same for hibernate, or to put the battery to power save mode.

So here is the simple way to do the same tasks from the cmd line.

pm : power management. check out the man pages for its various options.

pm-suspend : suspend/sleep the system

pm-hibernate : hibernate the system

pm-powersave : set the battery to power save mode

If you've lost your grub entries after installing a new os, here's a simple way of adding it back to your grub.Lets take the common situation where fedora has been installed along side ubuntu.In this case, the grub which you'l see is that of fedora, so you'l have to add the ubuntu entry into fedora's grub. first, boot into fedora and open the file '/boot/grub/menu.lst'.This is the file that has the entries which are visible when the grub is loaded. Amongst other things, the file would have a few statements that would look like

#################################### title Fedora (2.6.35.12-90.fc14.i686.PAE) root (hd0, 1) kernel /boot/vmlinuz-2.6.35.12-90.fc14.i686.PAE ro root=UUID=912d5b84-5c9c-4NO_LVM rd_NO_MD rd_NO_DM LANG=en_US.UTF-8 SYSFONT=latarcyrheb-sun16 KEYTABLE=us rhgb initrd /boot/initramfs-2.6.35.12-90.fc14.i686.PAE.img ###################################

these four lines are responsible for your grub to show the title 'fedora...' on the grub menu and boot it. To be short, the first line denotes the name to be displayed in the grub menu.The second line provides the hard disk and partiion (in which that particular os is existing) to be set as the root in the path. The third line is the full path to the kernel image of that os and the arguments that have to be passed for booting and the fourth line is the path to the initrd image

so if you want to add ubuntu(or any other) to the grub menu, all you have to do is to provide these parameters in the fedora's menu.lst file.These parameters will already be present in ubuntu(or any other) grub file.

mount your ubuntu partition and open the grub menu file(/boot/grub/menu.lst or grub.cfg).If the other os uses grub2 then you'l have to use grub.cfg instead of menu.lst Since ubuntu is now shipped with grub2, grub.cfg will have entries similar to

################################### menuentry 'Ubuntu,  with Linux 2.6.32-28-generic' --class ubuntu --class gnu-linux --class gnu --class os{ recordfail insmod ext2 set root='(hd0, 8)' search --no-floppy --fs-uuid --set f18af736-8d79-43e5-815e-0033f9c31ffe linux /boot/vmlinuz-2.6.32-28-generic root=UUID=f18af736-8d79-43e5-815e-0033f9c31ffe ro crashkernel=384M-2G:64M, 2G-:128M quiet splash initrd /boot/initrd.img-2.6.32-28-generic } ###################################

it has a differnt format, but contains the same information

1st line is the title

in the 4th line (hd0, 8) is the harddisk and the partition in that hard disk where the os is existing

6th line is the path to that os's kernel and its boot options7th line is the path to the initrd image of that os

so all you've to do now is to add these entries into fedora's menu.lst fileIn our case, the new entry in the menu.lst file would be

################################### title ubuntu root (hd0, 7) kernel /boot/vmlinuz-2.6.32-28-generic root=UUID=f18af736-8d79-43e5-815e-0033f9c31ffe ro crashkernel=384M-2G:64M, 2G-:128M quiet splash initrd /boot/initrd.img-2.6.32-28-generic ###################################

Note: in grub.cfg the partition was 8, but we use 7 in menu.lst because for this version of grub the partition number starts from 0