Skip to main content

Programmer's Guide: The "set" command in bash

The "set" command in shell allows you to set various shell options and positional parameters. The purpose of this post is to highlight how much this command has been helpful to me in my bash programming. I would not go into much depth about all the possible flag, because you can find many posts and documentation on the internet, like this post by GNU, which cover all the possible options that the "set" command provides. I will mainly focus on two options,

The "set -x" option: 

This option allows us to print commands and their arguments, after they are expanded (interpolated) but before they are executed.

Let us execute a simple hello world programme in bash.


When the programme is executed, we see the entire printf statement being printed before the actual "Hello World!" message is printed. This way, we can check the command that is being executed along with its output.


Setting the flag also expands any variables or arguments (interpolation) and prints them before execution. Furthermore, we can unset the option using "+" sign. 
In the below example, $1 implies that the programme should print 1st argument provided to it. Also, -x option is unset by setting +x.


If we run the code with 1st argument "Bash", observe that it prints Bash in the print statement before actually printing it. Also observe that the printf command is not shown after unsetting the -x option.

The "set -e" option:

This causes the programme to exit immediately if the command being executed returns a non-zero status. This is useful if you want to prevent further, possibly critical, code execution in case a certain command fails.

Continuing with our programme, I have modified the code to cat (print contents) of a file that doesn't exist. I have also "set -e".


The programme execution stops immediately and does not execute the critical section that comes after.


This flag is also useful in debugging a code. If you want to allow execution of certain commands which may cause errors but not allow certain commands whose output is not known, you can prevent termination of code by creating a complex command which returns status-code 0.

In the below code, we can continue execution in case a certain non-critical file doesn't exist, but we must not execute any further if a critical input file is not present.



As can be observed in the above execution, if a temporary file is missing and if we can safely ignore it, we can use "true" value and OR it with the command output to make a complex command whose exit status code is 0. You can read more about "true" value and exit status codes in bash here.
Status code 0 means that the overall command doesn't have any errors and the programme doesn't terminate. But it still terminates if a critical file is not present. So, execution of critical block is prevented which might have caused major issues.

Both these options have helped me tremendously while writing bash scripts. I hope you will find these options useful. Do explore other set options and let me know their uses in comments below.

Comments

  1. -x & -e are helpful flags while debugging shell scripts
    another approach is to use them in command line "bash -x script_name"

    ReplyDelete

Post a Comment

Popular posts from this blog

CRIF HighMark Credit Report Application Process

As per the RBI circular of 2017, you are entitled to get one free credit report every calendar year from each credit bureaus in India. CRIF HighMark is one of the four credit bureaus in India. This article will focus on creating a new account and getting your yearly free credit report. Visit the official website of CRIF HighMark https://www.crifhighmark.com/your-credit-score  In the next screen, click on " Get Your Score Now"  Register After registering, follow the below steps: Enter Personal Details Full Name DOB Email ID Mobile Number Gender Father's Name or Spouse Name Identification PAN or UID necessary Rest optional Communication Address CAPTCHA Review Terms of Use and Click Submit Choose free report Click " No Thanks, Take me to my FREE report " below the Upgrade me button. You will receive a confirmation email with your username and password and activation link. Click on activation link to...

My Entry into the World of Credit Cards

After joining my job in June 2018, my salary account was opened in ICICI Bank. During the account opening process, I was offered a Life Time Free (LTF) credit card. Simply put, this was my first chance to get into the credit world without any extra hassle because first of all, neither was I aware of the concept of credit scores nor I had any credit history and secondly, I was just curious about credit cards after seeing of my father use his credit cards with extra caution and the monthly event of credit card bill payment that used to happen at my home. And in that sudden moment of joy, I made my first mistake in the credit world. Representative Image (Source: https://jessepollak.github.io/card/) Now if your guess is that I started spending crazy amounts of money which I couldn't repay, then you are WRONG! I am a miserly person. Money doesn't leave my hands and in general, my credit utilization has never exceeded 5% of my credit limit. So coming back to my first mis...

Generate Large Files in Linux using dd

The programming community often requires large files for stress testing programmes. For example, sometimes such files are required to check the response time of certain programme or testing request handling capacity of servers. Many times, it does not matter what the contents of the file are, however, it is often difficult to find such large files when the need arises. Linux provides us a fast, efficient way of generating such huge files through simple command line options. For the purposes of this demo, I am using Ubuntu 16.04 and 18.04 linux distros. Ubuntu provides dd command to create such huge files in a matter of seconds. A typical dd command to create a 1 GB file is given below. dd if=/dev/zero of=big_file.txt count=1024 bs=1048576 parameters of dd: if - input file from which the content is read of - output file where the content is written count - number of blocks in output file bs - number of bytes in each block The above command creates a file named big_file.txt...