Trapping input from user [duplicate]

Instead of read -s -r -N 1 use read -e -s -r -N 1 call. Here only addition is the -s switch which tells read to use readline library for input and then only read will be able to understand arrow keys and other complex keystrokes.

The Root Cause: Arrow Keys Send Sequences

Here’s the deal: arrow keys aren’t “just a key.”
When you hit ←, you’re not sending one character — you’re sending a whole escape sequence.

For the usual arrows, it looks like this:

• Left: ESC [ D

• Right: ESC [ C

• Up: ESC [ A

• Down: ESC [ B

Now, your script is doing read -n1, which means “give me one char at a time.”
So when you press ←, this happens:

1. First loop: it grabs ESC. Doesn’t match anything.

2. Second loop: it grabs [. Still nothing.

3. Third loop: it grabs D. Bingo, that matches your d|D case. Script goes “oh cool, D was pressed,” even though you just wanted to move left.

That’s why it looks like the left arrow is pretending to be d.

The Fix: Eat the Whole Sequence

You’ve got to trap that first ESC, and then slurp up the next couple of chars so they don’t pollute your logic. Here’s a clean way to do it:

printf "Press$red Y$green to continue,$red N$green to Abort,$red D$green for Dry Run or$red H$green for Help\n\n" while true; do # using -n instead of -N read -s -r -n 1 Selection case "$Selection" in y | Y) echo "Y was pressed";; n | N) echo "N was pressed";; d | D) echo "D was pressed";; h | H) echo "H was pressed";; $'\e') # Arrow keys start with ESC, so grab the rest read -s -r -n 2 -t 0.1 rest;; # Do nothing — just flush it esac done 

$'\e' catches the start of any funky key (arrows, F-keys, etc)

read -n2 -t0.1 slurps up the [D part so it doesn’t leak into your loop