#!/usr/bin/perl
# natsume6.pl - Don Yang (uguu.org)
#
# 01/02/06


use Digest::MD5;

sub X
{
   ($v, $w) = @_;

   if( open F, "<$v" )
   {
      $q = Digest::MD5->new;
      if( $w )
      {
         $q->addfile(*F);
         $o += $w;
      }
      else
      {
         read F, $r, 1024;
         $q->add($r);
         $o += length $r;
      }
      close F;
      $r = $q->digest;
   }
   else
   {
      print "# $v: Can not open: $!\n";
      $r = undef;
   }
   return $r;
}

@r = @ARGV;
@r = <STDIN> if $#ARGV < 0;
chomp @r;

foreach $x (@r)
{
   @p = split /\//, $x;
   next if $#p < 0;
   @q = ();
   for($i = 0; $i <= $#p; $i++)
   {
      next if $p[$i] eq ".";

      if( $#q < 0 )
      {
         push @q, $p[$i];
      }
      else
      {
         if( $p[$i] ne ".." )
         {
            push @q, $p[$i];
         }
         elsif( $q[$#q] eq "" ||
                $q[$#q] eq ".." )
         {
            push @q, $p[$i];
         }
         else
         {
            pop @q;
         }
      }
   }

   $x = join '/', @q;
   push @l, $x if $x ne '' && $x ne '.' && $x ne '..';
}

if( $#l > 0 )
{
   @q = sort @l;
   @l = ($q[0]);
   foreach $x (@q)
   {
      push @l, $x if $x ne $l[$#l];
   }
}

$k = $o = $z = $u = $e = 0;
foreach $x (@l)
{
   if( -f $x && -r _ )
   {
      $b = -s _;
      $k++;
      $z += $b;
      if( $b <= 0 )
      {
         print "ln -s -f /dev/null '$x'\n";
         next;
      }

      if( !exists $a{$b} )
      {
         $a{$b}{0} = $x;
         next;
      }

      if( exists $a{$b}{0} )
      {
         $f = $a{$b}{0};
         X($f, 0);
         delete $a{$b}{0};
         $a{$b}{$r}{0} = $f;
      }

      $c = X($x, 0);
      if( !exists $a{$b}{$c} )
      {
         $a{$b}{$c}{0} = $x;
         next;
      }

      if( exists $a{$b}{$c}{0} )
      {
         $f = $a{$b}{$c}{0};
         X($f, -s $f);
         delete $a{$b}{$c}{0};
         $a{$b}{$c}{$r} = $f;
      }

      X($x, $b);
      if( !exists $a{$b}{$c}{$r} )
      {
         $a{$b}{$c}{$r} = $x;
         next;
      }
      $r = $a{$b}{$c}{$r};

      if( $r !~ m{^/} && $x !~ m{^/} )
      {
         $q = $x;
         while( $r =~ m{^([^/]+)/(.*)} )
         {
            $v = $1;
            $w = $2;
            last if $q !~ m{^([^/]+)/(.*)};
            last if $v ne $1;
            $r = $w;
            $q = $2;
         }
         if( index($q, '/') + 1 )
         {
            @p = split /\//, $q;
            $r = ("../" x $#p) . $r;
         }
      }

      print "ln -s -f '$r' '$x'\n";

      $u++;
      $e += $b;
   }
   else
   {
      print "# $x: not readable\n";
   }
}

print "# $k files, $o/$z bytes read\n",
   ($u > 0
      ? "# $e bytes in $u duplicate files\n"
      : "# No duplicates found\n");