#!/usr/bin/perl -w
use strict;
use File::Find::Duplicates;
use File::Temp ();

my %stats = ( files_linked => 0, space_saved => 0 );
local $" = "\n";

# Read directory from command line, or default to current.
my $directory = $ARGV[0] || ".";

# Find duplicates recursively in such directory
my @dupes = find_duplicate_files($directory);

# For each set of duplicate files, create the hardlinks and save the
# information in the stats hash
foreach my $set (@dupes) {
    print $set->size, " bytes each:\n", "@{ $set->files }\n";
    my $original = shift @{ $set->files };

    my $number_linked = fuse( $original, \@{ $set->files } );

    $stats{files_linked} += $number_linked;
    $stats{space_saved}  += $number_linked * $set->size;

}

# Report the stats
print "Files linked: $stats{ files_linked }\n";
print "Space saved:  $stats{ space_saved } bytes\n";

sub fuse {

    # Replace duplicates with hard links and return the number
    # of links created.

    my $original   = shift;
    my $duplicates = shift;
    my $files_linked;

    foreach my $duplicate (@$duplicates) {

        # Step 1: link original to tempfile
        my $tempfile = File::Temp::tempnam( $directory, 'X' x 6 );
        link $original, $tempfile or next;

        # Step 2: move tempfile to duplicate
        unless ( rename $tempfile, $duplicate ) { next; }
        if ( -e $tempfile ) {
            unlink $tempfile
                or die "Couldn't delete temporary file $tempfile: $!";
        }
        ++$files_linked;
    }

    return $files_linked;

}