#!/bin/sh

# From a single byte in hexadecimal per line to lines ending with 0a
# (hex for '\n'). Ex: 61 62 63 0a
# Required to easily match (and remove) multi-byte characters.
regroup_lines() awk '
	BEGIN {
		line_start=1
	}

	{
		if (line_start == 1)
			line = $1;
		else
			line = line " " $1;

		line_start = 0;
		if ($1 == "0a") {
			print line;
			line_start = 1
		}
	}

	END {
		if (line_start == 0)
			print line
	}
	'

uppercase()                   tr "[a-z]" "[A-Z]"

# One column decimal to plain text.
from_dec()                    awk '{ printf ("%c", $1 + 0) }'

# Replace spaces by line returns, outputs a single column.
spaces_to_line_returns()      tr " " "\n"

# Convert input into hexadecimal and a single byte per line.
to_hex_one_column() { od -An -tx1 | awk '{for(i=1;i<=NF;i++){ print $i }}'; }

# One column hexa to one column decimal.
hex_to_dec() { { echo "obase=10;ibase=16;" ; cat ; } | bc ; }

# Reverse hexadecimal (with space separators) to original value.
from_hex() { spaces_to_line_returns | uppercase | hex_to_dec | from_dec; }

# Replace unbreakable spaces by '\ '.
replace_unbreakable_spaces()    sed "s/c2 a0/5c 20/g"

custom_mods(){
	to_hex_one_column | # Input to hexadecimal, 1-byte representation per line.
		regroup_lines | # From 1-byte to x-byte lines with space separators.
		replace_unbreakable_spaces | # Replace unbreakable spaces by '\ '.
		from_hex # Convert back from hex (x-byte per line, space separator).
}

custom_mods