Testing user input arguments to a bash function
By Sarah Rodriguez •
I am handling the parsing of bash function arguments in the following way. The function affirm-numeric -W "$arg" tests if arg is a whole number (0,1,2,...), returning a value greater than zero, when arg is confirmed to be a whole number.
There are two possibilities to test the user inputs. As I do here, for each option. Otherwise make the checks after parsing the values through the while loop.
I wonder which would be the better approach in terms of functionality. It seems to me that the way I am doing it gives me more flexibility upon the parsing procedure.
while (( $# > 0 )); do case $1 in ("-T"|"--tail") # -T 21 nchk=$( affirm-numeric -W "$2" ) if (( nchk > 0 )); then tm="$2" ; shift ; shift else err+=(": -T failed numeric validation") fi ;; ("-T="*|"--tail="*) # -T=21 nchk=$( affirm-numeric -W "${1#*=}" ) if (( nchk > 0 )); then tm="${1#*=}" ; shift else err+=(": -T failed numeric validation") fi ;; ("-T"*) # -T21 nchk=$( affirm-numeric -W "${1#-T}" ) if (( nchk > 0 )); then tm="${1#*T}" ; shift else err+=(": -T failed numeric validation") fi ;; ("--") shift ; break ;; ("-"*) err=( "?" ) ; shift ;; (*) break ;; esac done
