Group
Extension

Net-RDAP/lib/Net/RDAP/Registry.pm

package Net::RDAP::Registry;
use Carp qw(croak);
use File::Basename qw(basename);
use File::Slurp;
use File::Spec;
use File::stat;
use JSON;
use POSIX qw(getpwuid);
use Net::RDAP::UA;
use Net::RDAP::Registry::IANARegistry;
use vars qw($UA $REGISTRY);
use constant {
    IP4_URL             => 'https://data.iana.org/rdap/ipv4.json',
    IP6_URL             => 'https://data.iana.org/rdap/ipv6.json',
    DNS_URL             => 'https://data.iana.org/rdap/dns.json',
    ASN_URL             => 'https://data.iana.org/rdap/asn.json',
    TAG_URL             => 'https://data.iana.org/rdap/object-tags.json',
    CACHE_TTL           => 86400,
};
use strict;

our $UA;

our $REGISTRY = {};

=pod

=head1 NAME

L<Net::RDAP::Registry> - a module which provides an interface to the IANA RDAP
registries.

=head1 SYNOPSIS

    use Net::RDAP::Registry;
    use Net::IP;
    use Net::ASN;

    $url = Net::RDAP::Registry->get_url(Net::DNS::Domain->new('example.com'));
    $url = Net::RDAP::Registry->get_url(Net::IP->new('192.168.0.1'));
    $url = Net::RDAP::Registry->get_url(Net::IP->new('2001:DB8::/32'));
    $url = Net::RDAP::Registry->get_url(Net::ASN->new(65536));
    $url = Net::RDAP::Registry->get_url('ABC123-TAG');

=head1 DESCRIPTION

RFC 7484 describes how RDAP clients can find the authoritative RDAP
service for a given internet resource using one of several IANA
registries.

This module provides an interface to these registries, and will return
a L<URI> object corresponding to the URL for a resource obtained from
them.

L<Net::RDAP::Registry> downloads the registry files from the IANA
website and will maintain up-to-date copies of those files locally.

=head1 METHODS

    $url = Net::RDAP::Registry->get_url($resource);

This method returns a L<URI> object corresponding to the authoritative
RDAP URL for the given resource. C<$resource> may be one of the
following:

=over

=item * a L<Net::IP> object representing an IPv4 or IPv6 address or
address range;

=item * a L<Net::ASN> object representing an Autonymous System;

=item * a L<Net::DNS::Domain> object representing a domain name.

=item * a string containing a tagged entity handle

=back

This method requires objects to be passed to ensure that the resource
identifiers have been properly validated.

If no URL can be found in the IANA registry, then C<undef> is returned.

=cut

sub get_url {
    my ($package, $object) = @_;

    if ('Net::IP' eq ref($object)) {
        return $package->ip($object);

    } elsif ('Net::ASN' eq ref($object)) {
        return $package->autnum($object);

    } elsif ('Net::DNS::Domain' eq ref($object)) {
        return $package->domain($object);

    } elsif ($object =~ /-/) {
        return $package->entity($object);

    } else {
        croak("Unable to deal with '$object'");

    }
}

#
# get URL for IP
#
sub ip {
    my ($package, $ip) = @_;
    croak(sprintf('Argument to %s->ip() must be a Net::IP', $package)) unless ('Net::IP' eq ref($ip));

    my $registry = $package->load_registry(4 == $ip->version ? IP4_URL : IP6_URL);
    return undef if (!$registry);

    my %matches;
    SERVICE: foreach my $service ($registry->services) {
        VALUE: foreach my $value ($service->registries) {
            my $range = Net::IP->new($value);

            if ($range->overlaps($ip)) {
                $matches{$value} = $package->get_best_url($service->urls);
                last VALUE;
            }
        }
    }

    return undef if (scalar(keys(%matches)) < 1);

    # prefer the service with the longest prefix length
    my $longest = (sort { Net::IP->new($b)->prefixlen <=> Net::IP->new($a)->prefixlen } keys(%matches))[0];

    return $package->assemble_url($matches{$longest}, 'ip', split(/\//, $ip->prefix));
}

#
# get URL for AS Number
#
sub autnum {
    my ($package, $autnum) = @_;
    croak(sprintf('Argument to %s->autnum() must be a Net::ASN', $package)) unless ('Net::ASN' eq ref($autnum));

    my $registry = $package->load_registry(ASN_URL);
    return undef if (!$registry);

    my %matches;
    SERVICE: foreach my $service ($registry->services) {
        VALUE: foreach my $value ($service->registries) {
            if ($value =~ /^\d+$/ && $value == $autnum->toasplain) {
                # exact match, create an entry for NNNN-NNNN where both sides
                # are the same (simplifies sorting later)
                $matches{sprintf('%d-%d', $value, $value)} = $package->get_best_url($service->urls);
                last SERVICE;

            } elsif ($value =~ /^(\d+)-(\d+)$/) {
                if ($1 <= $autnum->toasplain && $autnum->toasplain <= $2) {
                    $matches{$value} = $package->get_best_url($service->urls);
                    last VALUE;
                }
            }
        }
    }

    return undef if (scalar(keys(%matches)) < 1);

    my @ranges = keys(%matches);

    # convert array of NNNN-NNNN strings to array of array refs
    my @pairs = map { [ split(/-/, $_, 2) ] } @ranges;

    # sort by descending order of the "width" of the range
    my @sorted = sort { $b->{1} - $b->{0} <=> $a->{1} - $a->{0} } @pairs;

    # prefer the narrowest (more specific) range
    my $closest = sprintf('%d-%d', @{$sorted[0]});

    return $package->assemble_url($matches{$closest}, 'autnum', $autnum->toasplain);
}

#
# get URL for domain
#
sub domain {
    my ($package, $domain) = @_;
    croak(sprintf('Argument to %s->domain() must be a Net::DNS::Domain', $package)) unless ('Net::DNS::Domain' eq ref($domain));

    my $is_tld = (1 == scalar($domain->label));

    my $registry = $package->load_registry(DNS_URL);
    return undef if (!$registry);

    my %matches;
    SERVICE: foreach my $service ($registry->services) {
        VALUE: foreach my $value ($service->registries) {

            if ($is_tld && '' eq $value) {
                #
                # the RDAP server for the root zone is identified by an empty
                # string (see RFC 9224 ยง 4):
                #
                $matches{$value} = $package->get_best_url($service->urls);

                last VALUE;

            } else {
                if ($domain->name =~ /\.$value$/i) {
                    $matches{$value} = $package->get_best_url($service->urls);

                    last VALUE;
                }
            }
        }
    }

    if (scalar(keys(%matches)) < 1) {
        if ($domain->name =~ /\.(in-addr|ip6)\.arpa$/) {
            # special workaround for the lack of .arpa in the RDAP registry
            return $package->reverse_domain($domain);

        } else {
            return undef;

        }

    } else {
        # prefer the service with the longest domain name
        my $parent = (sort { length($b) <=> length($a) } keys(%matches))[0];

        return $package->assemble_url($matches{$parent}, 'domain', $domain->name);
    }
}

#
# construct the RDAP URL for a reverse domain. as of writing there's
# nothing in the IANA registry for the reverse tree so we work around
# that by by constructing the CIDR prefix that corresponds to the
# domain, resolving the RDAP URL for that, and then munging it to
# obtain the URL for the domain. clever, eh?
#
sub reverse_domain {
    my ($package, $domain) = @_;

    my @labels = reverse($domain->label);
    shift(@labels); # discard 'arpa'

    my $ip;
    if ('ip6' eq shift(@labels)) {
        # @labels is an array of hex digits, we want an array of 4-hex digit parts
        my @parts;
        push(@parts, join('', splice(@labels, 0, 4))) while (scalar(@labels) > 0);

        # remove any trailing parts that are zero
        pop(@parts) while (0 == hex($parts[-1]));

        # compute prefix length
        my $prefixlen = 16 * (scalar(@parts));

        $ip = Net::IP->new(sprintf(
            '%s:%s:%s:%s:%s:%s:%s:%s/%u',
            shift(@parts) || 0,
            shift(@parts) || 0,
            shift(@parts) || 0,
            shift(@parts) || 0,
            shift(@parts) || 0,
            shift(@parts) || 0,
            shift(@parts) || 0,
            shift(@parts) || 0,
            $prefixlen,
        ));

    } else {
        pop(@labels) while (0 == $labels[-1]);

        my $prefixlen = 8 * (scalar(@labels));

        $ip = Net::IP->new(sprintf(
            '%u.%u.%u.%u/%u',
            shift(@labels) || 0,
            shift(@labels) || 0,
            shift(@labels) || 0,
            shift(@labels) || 0,
            $prefixlen,
        ));
    }

    return undef if (!$ip);

    my $url = $package->ip($ip);

    return undef if (!$url);

    return URI->new_abs(sprintf('../../domain/%s', $domain->name), $url);
}

#
# get URL for a tagged entity
#
sub entity {
    my ($package, $handle) = @_;

    my @parts = split(/-/, $handle);
    my $tag = pop(@parts);

    my $registry = $package->load_registry(TAG_URL);
    return undef if (!$registry);

    foreach my $service ($registry->services) {
        foreach my $value ($service->registries) {
            # unlike the other registries we are only looking for an exact match, as there is no hierarchy:
            return $package->assemble_url($package->get_best_url($service->urls), 'entity', $handle) if (lc($value) eq lc($tag));
        }
    }

    return undef;
}

#
# load a registry. uses (in order of preference) an in-memory cache, a JSON file on disk,
# or a resource on the IANA website. returns a Net::RDAP::Registry::IANARegistry object
# (or undef)
#
sub load_registry {
    my ($package, $url) = @_;

    if (!defined($REGISTRY->{$url})) {
        $package =~ s/:+/-/g;

        my $file = sprintf(
            '%s/%s-%s-%s',
            File::Spec->tmpdir,
            $package,
            basename($url),
            getpwuid($<),
        );

        #
        # $UA may have been injected by Net::RDAP->ua()
        #
        $UA = Net::RDAP::UA->new unless(defined($UA));

        $UA->mirror($url, $file, CACHE_TTL);

        $REGISTRY->{$url} = Net::RDAP::Registry::IANARegistry->new(from_json(read_file($file))) if (-e $file);
    }

    return $REGISTRY->{$url};
}

#
# RDAP services can have multiple URLs, we pick the best by
# simply preferring the first one with the "https" scheme.
#
sub get_best_url {
    my ($package, @urls) = @_;

    my @https = grep { 'https' eq lc($_->scheme) } @urls;

    if (scalar(@https)) {
        return shift(@https);

    } else {
        return shift(@urls);

    }
}

#
# concatenate a URL with a bunch of path segments
# this method deals with URI objects which have
# trailing slashes
#
sub assemble_url {
    my ($package, $url, @segments) = @_;

    $url->path_segments(grep { defined && length > 0 } ($url->path_segments, @segments));

    return $url;
}

1;

__END__

=pod

=head1 COPYRIGHT

Copyright 2018-2023 CentralNic Ltd, 2024-2025 Gavin Brown. For licensing information,
please see the C<LICENSE> file in the L<Net::RDAP> distribution.

=cut


Powered by Groonga
Maintained by Kenichi Ishigaki <ishigaki@cpan.org>. If you find anything, submit it on GitHub.