#!/usr/bin/env perl
## Copyright © 2010 by Daniel Friesel <derf@finalrewind.org>
## License: WTFPL <http://sam.zoy.org/wtfpl>
##   0. You just DO WHAT THE FUCK YOU WANT TO.
use strict;
use warnings;
use 5.010;
use autodie;

use Cwd;
use Digest::SHA qw(sha1_hex);
use File::Find;
use Getopt::Long;
use Storable qw(nstore retrieve);
use Time::Progress;

my $base = getcwd();
my $rel_paths = 1;
my $read_size = (2 ** 20) * 4; # 4 MiB
my $db_file = "$base/.hashl.db";
my $total = 0;
my $cur = 0;
my $timer;

my $db;

GetOptions(
	'd|database=s' => \$db_file,
);

my $action = $ARGV[0];

if (not defined $action) {
	die("Usage: $0 <action>\n");
}

if (-r $db_file) {
	$db = retrieve($db_file);
}

sub get_total {
	my $file = $File::Find::name;
	if (-f $file and $file ne $db_file) {
		$total++;
	}
}

sub drop_deleted {
	for my $file (keys %{$db}) {
		if (! -e $file) {
			delete $db->{$file};
		}
	}
}

sub hash_file {
	my ($file) = @_;
	my ($fh, $data);

	open($fh, '<', $file);
	binmode($fh);
	read($fh, $data, $read_size);
	close($fh);

	return sha1_hex($data);
}

sub is_in_list {
	my ($file) = @_;
	my $hash = hash_file($file);

	if (grep { $_->{'hash'} eq $hash } values %{$db}) {
		return 1;
	}
	return 0;
}


sub process_file {
	my $file = $File::Find::name;
	my $path = $file;
	my ($size, $mtime) = (stat($file))[7,9];

	local $| = 1;

	if (not -f $file or $file eq $db_file) {
		return;
	}

	$cur++;

	if ($rel_paths) {
		$file = substr($file, length($base) + 1);
	}

	print $timer->report("\r\e[2KUpdating: %p done, %L elapsed, %E remaining", $cur);

	if (exists($db->{$file}) and
			$db->{$file}->{'mtime'} == $mtime and
			$db->{$file}->{'size'} == $size ) {
		return;
	}

	$db->{$file} = {
		hash => hash_file($path),
		mtime => $mtime,
		size => $size,
	};

	if (($cur % 100) == 0) {
		nstore($db, $db_file);
	}
}

if ($action eq 'update') {

	drop_deleted();
	find(\&get_total, $base);

	$timer = Time::Progress->new();
	$timer->attr(
		min => 1,
		max => $total
	);

	find(\&process_file, $base);
	print "\n";
	nstore($db, $db_file);
}
elsif ($action eq 'list') {
	for my $name (sort keys %{$db}) {
		my $file = $db->{$name};
		printf("%s %s\n", $file->{'hash'}, $name);
	}
}
elsif ($action eq 'in-list') {
	if ($ARGV[1]) {
		exit (!is_in_list($ARGV[1]));
	}
	else {
		while (my $line = <STDIN>) {
			chomp $line;
			if (!is_in_list($line)) {
				say $line;
			}
		}
	}
}

__END__

=head1 NAME

=head1 SYNOPSIS

=head1 DESCRIPTION

=head1 OPTIONS

=head1 EXIT STATUS

=head1 CONFIGURATION

=head1 DEPENDENCIES

=head1 BUGS AND LIMITATIONS

=head1 AUTHOR

Copyright (C) 2010 by Daniel Friesel E<lt>derf@finalrewind.orgE<gt>

=head1 LICENSE

  0. You just DO WHAT THE FUCK YOU WANT TO.