Bash Scripting

Abstract

There isn’t true or false
There is only exit code 0 means success, anything else is failure
A script has a return code of its last command which can be examined using echo $?

if [ -z "$a" ]; then
	echo "hello"
fi
 
a="123"
 
if [ -z "$a" ]; then
	echo "hello"
fi

By default, bash continues to execute the script if one of the command finished and didn’t exit with 0
The bash script will exit with 0, as long as the last command exits with 0
This is undesirable, because it create an illusion that all commands in the script managed to run successfully
The solution is to define set -e right below #!/bin/bash, so the bash script exit immediately if any command it performs fails (returns a non-zero status). Preventing it from continuing if something goes wrong
However, if you want the bash script to terminate when one of the commands in Pipe (管道) fails. You need to also add in set -o pipefail

We can place set -x to enable debugging info for the bash script
The shell will print each command before it’s executed, preceded by a ’+’ sign
This is useful for debugging scripts, as it allows you to see exactly what commands are being executed and spot any issues

By default, undefined bash variables are empty strings. They will not terminate the bash script!
We can add set -u to avoid this quirk

Keyword return is for Bash Script Exit Code, not return value
For return value, use echo instead
And use $(<BASH_FUNCTION_HERE>) command substitution to capture the return value

By default, the temporary changes like setting Environment Variable or directory changes is going to be removed once the bash script finishes
We can run the bash script with source which executes commands from a bash script in the current shell, so temporary changes are preserved after the bash script finishes running