code4Nix | Extract code blocks with ease

Many times, I come across scenario, wherein I need to split given file into sections and write it into different files or process them dynamically, as and when those sections are found.

I did solve this problem, although I find it to be rather inefficient, so I though of rewriting and came up with following generic function, which could be used by anyone to work on scenarios like extract function blocks from shell/perl/python scripts, or extract diff block from git diff output, etc.

Function Specification

function fn__extrt_block {

  test $# -ne 2 && exit 1
  test -f $2 && fv_iputFile=$2 || exit 2
  fv_srchPtrn="$1"

  fv_funcWorkDir=$HOME/${0}
  mkdir -p ${fv_funcWorkDir} && cd ${fv_funcWorkDir}

  grep -n "${fv_srchPtrn}" ${fv_iputFile} | cut -d':' -f1 > outfile

This grep command helps in finding out the line number to mark the beginning of block, using which following while loop will extract separate blocks.

  while true
  do
    echo $(head -2 outfile) | read LB UB
    test ! -z ${UB} && UB=`echo ${UB} - 1 | bc`

LB and UB represents the lower and upper bound of block, i.e. beginning and ending of specific block. For instance, if the outfile has contents like this;

1:Hi Varun

15:Hi Nischal

24:Hi Team

Every iteration, reads first two lines and assigns LB to 1, UB to 15. Next line of code, decrements UB by 1, so as to mark the end of block correctly.

    echo "${LB},${UB:-\$}p" > sed_scpt
    sed -n -f sed_scpt ${fv_iputFile} > file__${LB}_${UB:-$}

Once LB and UB are set, it becomes easy to extract block from input file using sed -n, as shown above. To continue iterating, it is important to keep removing first line, after every successful iteration.

    sed '1d' outfile > outfile.n
    mv outfile.n outfile

As this is an infinite loop, it is important to break the loop, once input file is completely processed.

    test ! -s outfile && break
  done
}

Usage

Let us say, you generate diff between two git commits using following command;

$> git diff versOne versTwo > ~/output__versOne_versTwo.diff

Then, execute the earlier defined function as follows;

$> fn__extrt_block '^diff' ~/output__versOne_versTwo.diff

This will generate block-specific files, with pattern file__*

GitHub Gist Code Reference

https://gist.github.com/nvarun/9575155

Hope this helps.

Advertisements

taT4Nix | Setting up Git repository in Fedora 12

I have been using Git for quite sometime now, although at work place. I could explore as much as I could, by setting up local repositories only. One day, I thought of cloning one of the repositories hosted at GitHub. However, I could not, possibly due to restrictions in place by System Administrator (I guess).

Setup GitHub Account and Installing Git

Now, to get more out of GitHub and explore further, I need a GitHub account and Linux OS. So I got started by setting up my account at GitHub and created repository, PrayogShala.

Recently, I had installed Fedora 12 (64 Bit) as Guest OS (via Virtual Box). So, I started the Virtual Machine and logged in with user account [nvarun] (created while installing the Guest OS) and executed steps as follows.

To begin contributing to repository at GitHub, one should install Git first and to do that, open up Terminal Desktop Application and refer yum steps in the reference guide.

Once dependencies are installed/updated, execute following command;

$> sudo yum install git

Once completed, find out the installed version of Git;

$> git --version
git version 1.7.2.3

Clone Repository

After installing Git, I cloned the repository earlier created on GitHub as follows;

$> git clone https://github.com/nvarun/PrayogShala.git PrayogShala

Then, based on my experiences at work place, I added useful git aliases in $HOME/.profile and started added/updating files in the repository. The interesting thing to notice, that all the changes I make, affects repository locally. To push the changes to remote location, i.e. the repository hosted at GitHub, execute following command;

$> git push origin master

Once you do that, the Terminal prompts for username and password of GitHub account. If validated with success, the changes are pushed successfully onto GitHub.

taT4Nix | Setting up Korn Shell in Linux OS

Having spent over a year at work place, working on AIX Server, I end up doing overtime writing tools using shell scripting in Korn Shell environment. I was wondering if there is some way possible, so that I could setup similar environment at home using Linux Desktop, like Ubuntu, Fedora, etc and avoid doing overtime 🙂

AIX Server uses ksh as the default shell for every user account, whereas most of the Linux OS (I guess) have bash as the default shell. As AIX is proprietary server made by IBM, so I opted for Virtual Box instead, where I could do all my exploration.

Why Fedora 12?

Just download the ISO Image for suitable flavor of Linux OS available in the Open Source community. Thankfully, I had Fedora 12 ISO Image, which I downloaded back in 2010. To save time, I used the same to setup VM using Virtual Box. 

While installing the Fedora 12 VM (on Windows 7 Host OS), I created user account [nvarun], which by default has bash shell environment loaded whenever I open the Terminal desktop application to do command line operations.

Setting up Korn Shell

I have spent past few weeks, trying to figure out how to change default shell for user account and loading ksh environment, whenever I open Terminal desktop application and user account [nvarun] logs in by default.

  1. Open Terminal desktop application
  2. By default, the user account [nvarun] is logged in.
  3. Execute following command;
    $> grep nvarun /etc/passwd
    nvarun:x:500:500:Varun Nischal:/home/nvarun:/bin/bash
  4. As you would notice, the entry is colon-delimited and the last delimited field tells you the default shell for user account.
  5. Now to change the default shell to /bin/ksh, execute following command;
    $> sudo chsh -s /bin/ksh nvarun
    Changing shell for nvarun.
    Shell changed.
    $> exit
  6. As you exit from Terminal and re-open it again, you won’t see the change yet.

Before we proceed further, it is important to understand what AIX Server does with user accounts.

The shell uses two types of profile files when you log in to the operating system.

The shell evaluates the commands contained in the files and then runs the commands to set up your system environment.

The files have similar functions except that the /etc/profile file controls profile variables for all users on a system whereas the .profile file allows you to customize your own environment.

As we need to customize the our environment, in that case, use vi editor to edit the $HOME/.profile file and add following line;

export ENV=/etc/kshrc

The /etc/kshrc file is a shell script that customizes the Korn Shell environment and often contains list of environment variables, aliases and function definitions, etc.

Each time you start a new instance of the Korn shell, the ksh command examines the value of the ENV environment variable set in the $HOME/.profile file.

If the ENV environment variable contains the name of an existing, readable file, the ksh command runs this file as a script.

By convention, this file is named $HOME/.kshrc. You can use another name, but you must set the ENV environment variable to point to it.

Now, you may exit from Terminal and re-open it again, the changes would be reflected and this lets nvarun create aliases, functions, etc. to customize the environment further.

I am already excited to explore further, hope this helps 🙂

taT4Nix | Useful Git Aliases

Couple of months back, I mentioned that I started using Git at workplace, thanks to my colleague. So far, I have been creating local repositories at workplace and in the process, spent lot of time exploring as much as I can.

This made me use the command line extensively and I ended up writing 3-letter aliases as follows, which helped saving lot of time.

alias gbr='git branch'
alias gco='git checkout'
alias gci='git commit'
alias gcm='git checkout master'
alias glo='git log --oneline'
alias gst='git status'

This would help, only if you are using command-line interfaces, on Linux/Unix Server.

taT4Nix | Transform HTML to WIKI using GNU sed

Recently, I have joined the forums – http://unix.stackexchange.com/ where I came across the following question;

http://unix.stackexchange.com/questions/11448/how-to-convert-a-href-http-xy-comxy-a-to-http-xy-com-xy

In simple words, this is what was required;

<a href="something">sometext</a>
# transforms into the following wiki markup
[something sometext]

Here’s what I feel should be done using GNU sed;

sed -r 's,<a\s+href=\"([^>]*)\">([^<]*)</a>,[\1 \2],g'

The curve-brackets helps in grouping pattern-matching data, which is then used-by back-references (eg. \1, \2, etc.)
So, in our scenario, we could identify two logical groups;

  1. Data between ‘<a href=”‘ and ‘”>’
  2. Data between ‘”>’ and ‘</a>’

Once groups are identified, using sed back-references, one can format it the way particular wiki markup requires.