#!/bin/sh
# Steve Parker, http://steve-parker.org/
# Updated Nov 2005 to work properly with filenames containing spaces
# Checks for duplicate files in a directory 
# tree and removes them (or does whatever else
# you want to do with them).
# Uses diff to check the difference, not filename.
# Depends on:
#    diff, sum

LOG=/tmp/sum.$$
trap 'echo "Interrupted... " ; rm -f ${LOG} ; exit 1' 1 2 15

touch ${LOG}

find . -type f -print | while read i 
do
  echo -n "."
  #echo "Testing $i ..."
  s=`sum "$i"|cut -d" " -f1`
  n=`grep "^${s}:" ${LOG}`
  #echo "s=$s n=$n"
  if [ ! -z "$n" ]; then
    for j in `echo $n | cut -d":" -f2-`
    do
      diff "$i" "$j" > /dev/null 2>&1
      if [ "$?" -eq "0" ]; then 
        # This is the bit you want to change.
        # $i is the current file, $j is a previous file
	# which has already been found, and is identical to $i.
	# examples:
        # echo "$i and $j match"
	# ls -l $i $j
	# rm -f $i
        echo "$i and $j match"
      fi
    done
  else
    echo "$s:${i}" >> ${LOG}
  fi
done
rm -f ${LOG}