#!/bin/bash
#===============================================================================
# calculates the difference between two files                             
#===============================================================================

# get command line arguments or print usage and die
[ "$1" = "-b" ] && { binary=true ; shift; }
[ $# -ne 2 ] && {
    echo >&2 "usage: $(basename "$0") [-b] file1 file2"
    echo >&2 "    -b    compare binary files instead of texts"
    exit 1
}
file1="$1"
file2="$2"

# prepare for a diff:
prepare() { 
    if [ -n "$binary" ]
    then hexdump -v -e '1/1 "%0.2x\n"'
    else tr -d ' \t\r\n' | sed 's/\(.\)/\1\n/g'
    fi
}

# load prepared input files into memory
content1=$(prepare <"$file1")
content2=$(prepare <"$file2")

# count number of characters in the prepares files content
count1=$(echo "$content1" | wc -l)
count2=$(echo "$content2" | wc -l)

# diff the prepared content, grep differences and count them
difference=$(
    diff <(echo "$content1") <(echo "$content2") |
    grep '^\(<\|>\)' | wc -l
)

# calculate sum of characters in both files
sum=$((count1 + count2))

# calculate percent of changes
percent=$((difference * 100 / sum))

# print results
echo "$difference/$sum $percent%"

#===============================================================================
#:mode=shellscript:noTabs=false:tabSize=4:indentSize=4:lineSeparator=\n: