Error Handling and Functions with BASH

BASH[1], to my understanding is not meant to be used as a general purpose programming language and attempting to use it for such purposes can lead to tricky situations. Having said that, for the sake of argument let’s assume you need to use it for such a purpose. Here's how you can do some of the housekeeping tasks with it.

Error Handling and API Calls

Process exit codes can be used to do a quick assessment of successful completion, however it should be kept in mind that checking the  error codes alone cannot guarantee output is error free and as you expect it to be.

“?” env variable contains the exit code for the last run process in session, where a 0 denotes a successful process exit while 1 denotes a failure. This variable when coupled with conditional statements can bring in some basic error handling capabilities to your scripts.

	#do something if error code denotes a failure
	if [ $? -gt 0 ]; then
	echo "something went wrong!"
	fi
	#status checking in command pipe
	cat textfile.txt | ( [[ $? == 0 ]] && grep "cat" )

view raw gistfile1.sh hosted with ❤ by GitHub

API calls can be made using the cURL HTTP client. This package comes installed in many Linux distributions but it's advisable to do a check for availability using the package management tools available in the system.

A simple way to validate API calls would be to utilize the -v flag in cURL client to print out a verbose output and then redirect the output to a grep to be asserted for a predefined value.

	curl -X GET -v $ENDPOINT > RESPONSE
	cat $RESPONSE | grep -i "success"

view raw gistfile1.sh hosted with ❤ by GitHub

Functions

Bash provides limited capability to abstract out code to create functions, find more about its capabilities with regard to functions here[2]. As BASH is not capable of creating user defined data types, you might find yourself in a situation where you need to return a bunch of data items back to the caller. A crude way in which this requirement can be achieved is by dividing the data items with a unique delimiter and then breaking up the values on the caller side using awk.

	findproductnames()
	{
	DELIMITER="*"
	PROD_ONE="Apache JMeter"
	PROD_TWO="Apache Tomcat"
	echo $PROD_ONE$DELIMITER$PROD_TWO
	return
	}
	FUNCTION_CALL_RETURN_VALUES="$(findproductnames)"
	echo $FUNCTION_CALL_RETURN_VALUES
	VALUE_ONE="$(echo $FUNCTION_CALL_RETURN_VALUES | awk -F '*' '{print $1}')"
	VALUE_TWO="$(echo $FUNCTION_CALL_RETURN_VALUES | awk -F '*' '{print $2}')"
	echo $VALUE_ONE$VALUE_TWO

view raw gistfile1.sh hosted with ❤ by GitHub

When you start resorting to elaborate methods to get simple tasks done, you are probably extending BASH beyond its capabilities and it’s time to move onto a more capable language such as python.


[1] - https://www.gnu.org/software/bash/manual/html_node/
[2] - http://www.linuxjournal.com/content/return-values-bash-functions