#!/bin/env ruby
#
# This does something similar as
#
# $ find <path> -type f -print0 | xargs -0 md5sum | sort | uniq --all-repeated=separate -w32
#
# but dumps a collection of ln command lines to execute.

require 'md5'

def checksums(dir)
  files = Dir[File.join(dir, "**", "*")].select {|name| !File.symlink?(name) && File.file?(name) }
  files.map {|fname| [fname, MD5.new(File.open(fname, "rb").read).to_s] }
end

def merger(*checksums)
  checksums.inject({}) do |all_checksums,file_checksums|
    file_checksums.each do |fname,sum|
      all_checksums.merge!({ sum => [fname] }) {|k,o,n| o | n }
    end
    all_checksums
  end
end

def hardlinker(merged_checksums)
  merged_checksums.each do |sum,files|
    next if files.count < 2
    src = files.shift
    files.each {|dst| puts "ln -fn '#{src}' '#{dst}'" }
  end
end