I have a Bash script that sets various environment variables and runs a program. I want one of those variables to change if I run the command with an argument debug.
I know 100% how to do this in Lua code, but Bash is not Lua and I'm still in the early stages of learning it.
Essentially, when I run ./foo.sh it does
export var1
export var2
export var3
program
I want it to not pass var1 if I run ./foo.sh debug
Change
#!/bin/bash
export var1="value1"
export var2="value2"
export var3="value3"
program
into
#!/bin/bash
if [ "$1" != "debug" ]; then
export var1="value1"
fi
export var2="value2"
export var3="value3"
program
The first command line argument to your script will be available in $1, and if that is not the string debug, export the var1 environment variable.
The #!-line could optionally be changed into #!/bin/sh as the script (as written above) does not use any bash-specific features.
A variation of the above which does it differently with an array, without actually setting the variables in the script itself (only for the program when starting it):
#!/bin/bash
if [ "$1" != "debug" ]; then
vars=( var1="value1" )
fi
vars+=( var2="value2" var3="value3" )
env "${vars[@]}" program
or, for /bin/sh,
#!/bin/sh
mode=$1
set -- var2="value2" var3="value3"
if [ "$mode" != "debug" ]; then
set -- var1="value1" "$@"
fi
env "$@" program
#!/bin/sh on all scripts that does not use bash-specific extensions such as arrays or the [[ ... ]]-like tests etc. I do this mostly because I think it looks neater but also for portability (some of my scripts must run on systems where bash is not available). You should never have to end a script with exit as the last line of the script. An exit 0 would also "mask" the real exit status of the script which would otherwise be that of the most recently executed command (program in your case here).
if [[ $1 != debug ]]; then export var1; fi