CIDR.pm - metacpan.org

# Net::CIDR
#
# Copyright 2001-2012 Sam Varshavchik.
#
# with contributions from David Cantrell.
#
# This program is free software; you can redistribute it
# and/or modify it under the same terms as Perl itself.

package Net::CIDR;

require 5.000;
#use strict;
#use warnings;

require Exporter;
# use AutoLoader qw(AUTOLOAD);
use Carp;

@ISA = qw(Exporter);

# Items to export into callers namespace by default. Note: do not export
# names by default without a very good reason. Use EXPORT_OK instead.
# Do not simply export all your public functions/methods/constants.

# This allows declaration	use Net::CIDR ':all';
# If you do not need this, moving things directly into @EXPORT or @EXPORT_OK
# will save memory.
%EXPORT_TAGS = ( 'all' => [ qw( range2cidr
				    cidr2range
				    cidr2octets
				    cidradd
				    cidrlookup
				    cidrvalidate
				    addr2cidr
                                    addrandmask2cidr
				    ) ] );

@EXPORT_OK = ( qw( range2cidr
		       cidr2range
		       cidr2octets
		       cidradd
		       cidrlookup
		       cidrvalidate
		       addr2cidr
                       addrandmask2cidr
		       ));

@EXPORT = qw(
	
);

$VERSION = "0.17";

1;


=pod

=head1 NAME

Net::CIDR - Manipulate IPv4/IPv6 netblocks in CIDR notation

=head1 SYNOPSIS

    use Net::CIDR;

    use Net::CIDR ':all';

    print join("\n",
          Net::CIDR::range2cidr("192.68.0.0-192.68.255.255",
		                "10.0.0.0-10.3.255.255"))
	       . "\n";
    #
    # Output from above:
    #
    # 192.68.0.0/16
    # 10.0.0.0/14

    print join("\n",
          Net::CIDR::range2cidr(
		"dead:beef::-dead:beef:ffff:ffff:ffff:ffff:ffff:ffff"))
               . "\n";

    #
    # Output from above:
    #
    # dead:beef::/32

    print join("\n",
	     Net::CIDR::range2cidr("192.68.1.0-192.68.2.255"))
                  . "\n";
    #
    # Output from above:
    #
    # 192.68.1.0/24
    # 192.68.2.0/24

    print join("\n", Net::CIDR::cidr2range("192.68.0.0/16")) . "\n";
    #
    # Output from above:
    #
    # 192.68.0.0-192.68.255.255

    print join("\n", Net::CIDR::cidr2range("dead::beef::/46")) . "\n";
    #
    # Output from above:
    #
    # dead:beef::-dead:beef:3:ffff:ffff:ffff:ffff:ffff

    @list=("192.68.0.0/24");
    @list=Net::CIDR::cidradd("192.68.1.0-192.68.1.255", @list);

    print join("\n", @list) . "\n";
    #
    # Output from above:
    #
    # 192.68.0.0/23

    print join("\n", Net::CIDR::cidr2octets("192.68.0.0/22")) . "\n";
    #
    # Output from above:
    #
    # 192.68.0
    # 192.68.1
    # 192.68.2
    # 192.68.3

    print join("\n", Net::CIDR::cidr2octets("dead::beef::/46")) . "\n";
    #
    # Output from above:
    #
    # dead:beef:0000
    # dead:beef:0001
    # dead:beef:0002
    # dead:beef:0003

    @list=("192.68.0.0/24");
    print Net::CIDR::cidrlookup("192.68.0.12", @list);
    #
    # Output from above:
    #
    # 1

    @list = Net::CIDR::addr2cidr("192.68.0.31");
    print join("\n", @list);
    #
    # Output from above:
    #
    # 192.68.0.31/32
    # 192.68.0.30/31
    # 192.68.0.28/30
    # 192.68.0.24/29
    # 192.68.0.16/28
    # 192.68.0.0/27
    # 192.68.0.0/26
    # 192.68.0.0/25
    # 192.68.0.0/24
    # 192.68.0.0/23
    # [and so on]

    print Net::CIDR::addrandmask2cidr("195.149.50.61", "255.255.255.248")."\n";
    #
    # Output from above:
    #
    # 195.149.50.56/29

=head1 DESCRIPTION

The Net::CIDR package contains functions that manipulate lists of IP
netblocks expressed in CIDR notation.
The Net::CIDR functions handle both IPv4 and IPv6 addresses.

=head2 @cidr_list=Net::CIDR::range2cidr(@range_list);

Each element in the @range_list is a string "start-finish", where
"start" is the first IP address and "finish" is the last IP address.
range2cidr() converts each range into an equivalent CIDR netblock.
It returns a list of netblocks except in the case where it is given
only one parameter and is called in scalar context.

For example:

    @a=Net::CIDR::range2cidr("192.68.0.0-192.68.255.255");

The result is a one-element array, with $a[0] being "192.68.0.0/16".
range2cidr() processes each "start-finish" element in @range_list separately.
But if invoked like so:

    $a=Net::CIDR::range2cidr("192.68.0.0-192.68.255.255");

The result is a scalar "192.68.0.0/16".

Where each element cannot be expressed as a single CIDR netblock
range2cidr() will generate as many CIDR netblocks as are necessary to cover
the full range of IP addresses.  Example:

    @a=Net::CIDR::range2cidr("192.68.1.0-192.68.2.255");

The result is a two element array: ("192.68.1.0/24","192.68.2.0/24");

    @a=Net::CIDR::range2cidr(
		   "d08c:43::-d08c:43:ffff:ffff:ffff:ffff:ffff:ffff");

The result is an one element array: ("d08c:43::/32") that reflects this
IPv6 netblock in CIDR notation.

range2cidr() does not merge adjacent or overlapping netblocks in
@range_list.

=head2 @range_list=Net::CIDR::cidr2range(@cidr_list);

The cidr2range() functions converts a netblock list in CIDR notation
to a list of "start-finish" IP address ranges:

    @a=Net::CIDR::cidr2range("10.0.0.0/14", "192.68.0.0/24");

The result is a two-element array: 
("10.0.0.0-10.3.255.255", "192.68.0.0-192.68.0.255").

    @a=Net::CIDR::cidr2range("d08c:43::/32");

The result is a one-element array:
("d08c:43::-d08c:43:ffff:ffff:ffff:ffff:ffff:ffff").

cidr2range() does not merge adjacent or overlapping netblocks in
@cidr_list.

=head2 @netblock_list = Net::CIDR::addr2cidr($address);

The addr2cidr function takes an IP address and returns a list of all
the CIDR netblocks it might belong to:

    @a=Net::CIDR::addr2cidr('192.68.0.31');

The result is a thirtythree-element array:
('192.68.0.31/32', '192.68.0.30/31', '192.68.0.28/30', '192.68.0.24/29',
 [and so on])
consisting of all the possible subnets containing this address from
0.0.0.0/0 to address/32.

Any addresses supplied to addr2cidr after the first will be ignored.
It works similarly for IPv6 addresses, returning a list of one hundred
and twenty nine elements.

=head2 $cidr=Net::CIDR::addrandmask2cidr($address, $netmask);

The addrandmask2cidr function takes an IP address and a netmask, and
returns the CIDR range whose size fits the netmask and which contains
the address.  It is an error to supply one parameter in IPv4-ish
format and the other in IPv6-ish format, and it is an error to supply
a netmask which does not consist solely of 1 bits followed by 0 bits.
For example, '255.255.248.192' is an invalid netmask, as is
'255.255.255.32' because both contain 0 bits in between 1 bits.

Technically speaking both of those *are* valid netmasks, but a) you'd
have to be insane to use them, and b) there's no corresponding CIDR
range.

=cut

# CIDR to start-finish

sub cidr2range {
    my @cidr=@_;

    my @r;

    while ($#cidr >= 0)
    {
	my $cidr=shift @cidr;

	$cidr =~ s/\s//g;

	unless ($cidr =~ /(.*)\/(.*)/)
	{
	    push @r, $cidr;
	    next;
	}

	my ($ip, $pfix)=($1, $2);

	my $isipv6;

	my @ips=_iptoipa($ip);

	$isipv6=shift @ips;

	croak "$pfix, as in '$cidr', does not make sense"
	    unless $pfix >= 0 && $pfix <= ($#ips+1) * 8 && $pfix =~ /^[0-9]+$/;

	my @rr=_cidr2iprange($pfix, @ips);

	while ($#rr >= 0)
	{
	    my $a=shift @rr;
	    my $b=shift @rr;

	    $a =~ s/\.$//;
	    $b =~ s/\.$//;

	    if ($isipv6)
	    {
		$a=_ipv4to6($a);
		$b=_ipv4to6($b);
	    }

	    push @r, "$a-$b";
	}
    }

    return @r;
}

#
# If the input is an IPv6-formatted address, convert it to an IPv4 decimal
# format, since the other functions know how to deal with it.  The hexadecimal
# IPv6 address is represented in dotted-decimal form, like IPv4.
#

sub _ipv6to4 {
    my $ipv6=shift;

    return (undef, $ipv6) unless $ipv6 =~ /:/;

    croak "Syntax error: $ipv6"
	unless $ipv6 =~ /^[a-fA-F0-9:\.]+$/;

    my $ip4_suffix="";

    ($ipv6, $ip4_suffix)=($1, $2)
	if $ipv6 =~ /^(.*:)([0-9]+\.[0-9\.]+)$/;

    $ipv6 =~ s/([a-fA-F0-9]+)/_h62d($1)/ge;

    my $ipv6_suffix="";

    if ($ipv6 =~ /(.*)::(.*)/)
    {
	($ipv6, $ipv6_suffix)=($1, $2);
	$ipv6_suffix .= ".$ip4_suffix";
    }
    else
    {
	$ipv6 .= ".$ip4_suffix";
    }

    my @p=grep (/./, split (/[^0-9]+/, $ipv6));

    my @s=grep (/./, split (/[^0-9]+/, $ipv6_suffix));

    push @p, 0 while $#p + $#s < 14;

    my $n=join(".", @p, @s);

#    return (undef, $1)
#	if $n =~ /^0\.0\.0\.0\.0\.0\.0\.0\.0\.0\.255\.255\.(.*)$/;

    return (1, $n);
}

# Let's go the other way around

sub _ipv4to6 {
    my @octets=split(/[^0-9]+/, shift);

    croak "Internal error in _ipv4to6"
	unless $#octets == 15;

    my @dummy=@octets;

    return ("::ffff:" . join(".", $octets[12], $octets[13], $octets[14], $octets[15]))
	if join(".", splice(@dummy, 0, 12)) eq "0.0.0.0.0.0.0.0.0.0.255.255";

    my @words;

    my $i;

    for ($i=0; $i < 8; $i++)
    {
	$words[$i]=sprintf("%x", $octets[$i*2] * 256 + $octets[$i*2+1]);
    }

    my $ind= -1;
    my $indlen= -1;

    for ($i=0; $i < 8; $i++)
    {
	next unless $words[$i] eq "0";

	my $j;

	for ($j=$i; $j < 8; $j++)
	{
	    last if $words[$j] ne "0";
	}

	if ($j - $i > $indlen)
	{
	    $indlen= $j-$i;
	    $ind=$i;
	    $i=$j-1;
	}
    }

    return "::" if $indlen == 8;

    return join(":", @words) if $ind < 0;

    my @s=splice (@words, $ind+$indlen);

    return join(":", splice (@words, 0, $ind)) . "::"
	. join(":", @s);
}

# An IP address to an octet list.

# Returns a list. First element, flag: true if it was an IPv6 flag. Remaining
# values are octets.

sub _iptoipa {
    my $iparg=shift;

    my $isipv6;
    my $ip;

    ($isipv6, $ip)=_ipv6to4($iparg);

    my @ips= split (/\.+/, $ip);

    grep {
	croak "$_, in $iparg, is not a byte" unless $_ >= 0 && $_ <= 255 && $_ =~ /^[0-9]+$/;
    } @ips;

    return ($isipv6, @ips);
}
    
sub _h62d {
    my $h=shift;

    $h=hex("0x$h");

    return ( int($h / 256) . "." . ($h % 256));
}

sub _cidr2iprange {
    my @ips=@_;
    my $pfix=shift @ips;

    if ($pfix == 0)
    {
	grep { $_=0 } @ips;

	my @ips2=@ips;

	grep { $_=255 } @ips2;

	return ( join(".", @ips), join(".", @ips2));
    }

    if ($pfix >= 8)
    {
	my $octet=shift @ips;

	@ips=_cidr2iprange($pfix - 8, @ips);

	grep { $_="$octet.$_"; } @ips;
	return @ips;
    }

    my $octet=shift @ips;

    grep { $_=0 } @ips;

    my @ips2=@ips;

    grep { $_=255 } @ips2;

    my @r= _cidr2range8(($octet, $pfix));

    $r[0] = join (".", ($r[0], @ips));
    $r[1] = join (".", ($r[1], @ips2));

    return @r;
}

#
# ADDRESS to list of CIDR netblocks
#

sub addr2cidr {
    my @ips=_iptoipa(shift);

    my $isipv6=shift @ips;

    my $nbits;

    if ($isipv6)
    {
	croak "An IPv6 address is 16 bytes long" unless $#ips == 15;
	$nbits=128;
    }
    else
    {
	croak "An IPv4 address is 4 bytes long" unless $#ips == 3;
	$nbits=32;
    }

    my @blocks;

    foreach my $bits (reverse 0..$nbits)
    {
	my @ipcpy=@ips;

	my $n=$bits;

	while ($n < $nbits)
	{
	    @ipcpy[$n / 8] &= (0xFF00 >> ($n % 8));

	    $n += 8;

	    $n &= 0xF8;
	}

	my $s=join(".", @ipcpy);

	push @blocks, ($isipv6 ? _ipv4to6($s):$s) . "/$bits";
    }
    return @blocks;
}

# Address and netmask to CIDR

sub addrandmask2cidr {
        my $address = shift;
	my($a_isIPv6) = _ipv6to4($address);
        my($n_isIPv6, $netmask) = _ipv6to4(shift);
	die("Both address and netmask must be the same type")
	    if( defined($a_isIPv6) && defined($n_isIPv6) && $a_isIPv6 != $n_isIPv6);
        my $bitsInNetmask = 0;
        my $previousNMoctet = 255;
        foreach my $octet (split/\./, $netmask) {
            die("Invalid netmask") if($previousNMoctet != 255 && $octet != 0);
            $previousNMoctet = $octet;
	    $bitsInNetmask +=
		($octet == 255) ? 8 :
		($octet == 254) ? 7 :
		($octet == 252) ? 6 :
		($octet == 248) ? 5 :
		($octet == 240) ? 4 :
		($octet == 224) ? 3 :
		($octet == 192) ? 2 :
		($octet == 128) ? 1 :
		($octet == 0) ? 0 :
                die("Invalid netmask");
	}
        return (grep { /\/$bitsInNetmask$/ } addr2cidr($address))[0];
}

#
# START-FINISH to CIDR list
#

sub range2cidr {
    my @r=@_;

    my $i;

    my @c;

    for ($i=0; $i <= $#r; $i++)
    {
	$r[$i] =~ s/\s//g;

	if ($r[$i] =~ /\//)
	{
	    push @c, $r[$i];
	    next;
	}

	$r[$i]="$r[$i]-$r[$i]" unless $r[$i] =~ /(.*)-(.*)/;

	$r[$i] =~ /(.*)-(.*)/;

	my ($a,$b)=($1,$2);

	my $isipv6_1;
	my $isipv6_2;

	($isipv6_1, $a)=_ipv6to4($a);
	($isipv6_2, $b)=_ipv6to4($b);

	if ($isipv6_1 || $isipv6_2)
	{
	    croak "Invalid netblock range: $r[$i]"
		unless $isipv6_1 && $isipv6_2;
	}

	my @a=split(/\.+/, $a);
	my @b=split(/\.+/, $b);

	croak unless $#a == $#b;

	my @cc=_range2cidr(\@a, \@b);

	while ($#cc >= 0)
	{
	    $a=shift @cc;
	    $b=shift @cc;

	    $a=_ipv4to6($a) if $isipv6_1;

	    push @c, "$a/$b";
	}
    }
    return @c unless(1==@r && 1==@c && !wantarray());
    return $c[0];
}

sub _range2cidr {
    my $a=shift;
    my $b=shift;

    my @a=@$a;
    my @b=@$b;

    $a=shift @a;
    $b=shift @b;

    return _range2cidr8($a, $b) if $#a < 0; # Least significant octet pair.

    croak "Bad starting address\n" unless $a >= 0 && $a <= 255 && $a =~ /^[0-9]+$/;
    croak "Bad ending address\n" unless $b >= 0 && $b <= 255 && $b =~ /^[0-9]+$/ && $b >= $a;

    my @c;

    if ($a == $b) # Same start/end octet
    {
	my @cc= _range2cidr(\@a, \@b);

	while ($#cc >= 0)
	{
	    my $c=shift @cc;

	    push @c, "$a.$c";

	    $c=shift @cc;
	    push @c, $c+8;
	}
	return @c;
    }

    my $start0=1;
    my $end255=1;

    grep { $start0=0 unless $_ == 0; } @a;
    grep { $end255=0 unless $_ == 255; } @b;

    if ( ! $start0 )
    {
	my @bcopy=@b;

	grep { $_=255 } @bcopy;

	my @cc= _range2cidr(\@a, \@bcopy);

	while ($#cc >= 0)
	{
	    my $c=shift @cc;

	    push @c, "$a.$c";

	    $c=shift @cc;
	    push @c, $c + 8;
	}

	++$a;
    }

    if ( ! $end255 )
    {
	my @acopy=@a;

	grep { $_=0 } @acopy;

	my @cc= _range2cidr(\@acopy, \@b);

	while ($#cc >= 0)
	{
	    my $c=shift @cc;

	    push @c, "$b.$c";

	    $c=shift @cc;
	    push @c, $c + 8;
	}

	--$b;
    }

    if ($a <= $b)
    {
	grep { $_=0 } @a;

	my $pfix=join(".", @a);

	my @cc= _range2cidr8($a, $b);

	while ($#cc >= 0)
	{
	    my $c=shift @cc;

	    push @c, "$c.$pfix";

	    $c=shift @cc;
	    push @c, $c;
	}
    }
    return @c;
}

sub _range2cidr8 {

    my @c;

    my @r=@_;

    while ($#r >= 0)
    {
	my $a=shift @r;
	my $b=shift @r;

	croak "Bad starting address\n" unless $a >= 0 && $a <= 255 && $a =~ /^[0-9]+$/;
	croak "Bad ending address\n" unless $b >= 0 && $b <= 255 && $b =~ /^[0-9]+$/ && $b >= $a;

	++$b;

	while ($a < $b)
	{
	    my $i=0;
	    my $n=1;

	    while ( ($n & $a) == 0)
	    {
		++$i;
		$n <<= 1;
		last if $i >= 8;
	    }

	    while ($i && $n + $a > $b)
	    {
		--$i;
		$n >>= 1;
	    }

	    push @c, $a;
	    push @c, 8-$i;

	    $a += $n;
	}
    }

    return @c;
}

sub _cidr2range8 {

    my @c=@_;

    my @r;

    while ($#c >= 0)
    {
	my $a=shift @c;
	my $b=shift @c;

	croak "Bad starting address" unless $a >= 0 && $a <= 255 && $a =~ /^[0-9]+$/;
	croak "Bad ending address" unless $b >= 0 && $b <= 8 && $b =~ /^[0-9]+$/;

	my $n= 1 << (8-$b);

	$a &= ($n-1) ^ 255;

	push @r, $a;
	push @r, $a + ($n-1);
    }
    return @r;
}

sub _ipcmp {
    my $aa=shift;
    my $bb=shift;

    my $isipv6_1;
    my $isipv6_2;

    ($isipv6_1, $aa)=_ipv6to4($aa);
    ($isipv6_2, $bb)=_ipv6to4($bb);

    if ($isipv6_1 || $isipv6_2)
    {
	croak "Invalid netblock: $aa-$bb"
	    unless $isipv6_1 && $isipv6_2;
    }

    my @a=split (/\./, $aa);
    my @b=split (/\./, $bb);

    croak "Different number of octets in IP addresses" unless $#a == $#b;

    while ($#a >= 0 && $a[0] == $b[0])
    {
	shift @a;
	shift @b;
    }

    return 0 if $#a < 0;

    return $a[0] <=> $b[0];
}


=pod

=head2 @octet_list=Net::CIDR::cidr2octets(@cidr_list);

cidr2octets() takes @cidr_list and returns a list of leading octets
representing those netblocks.  Example:

    @octet_list=Net::CIDR::cidr2octets("10.0.0.0/14", "192.68.0.0/24");

The result is the following five-element array:
("10.0", "10.1", "10.2", "10.3", "192.68.0").

For IPv6 addresses, the hexadecimal words in the resulting list are
zero-padded:

    @octet_list=Net::CIDR::cidr2octets("::dead:beef:0:0/110");

The result is a four-element array:
("0000:0000:0000:0000:dead:beef:0000",
"0000:0000:0000:0000:dead:beef:0001",
"0000:0000:0000:0000:dead:beef:0002",
"0000:0000:0000:0000:dead:beef:0003").
Prefixes of IPv6 CIDR blocks should be even multiples of 16 bits, otherwise
they can potentially expand out to a 32,768-element array, each!

=cut

sub cidr2octets {
    my @cidr=@_;

    my @r;

    while ($#cidr >= 0)
    {
	my $cidr=shift @cidr;

	$cidr =~ s/\s//g;

	croak "CIDR doesn't look like a CIDR\n" unless ($cidr =~ /(.*)\/(.*)/);

	my ($ip, $pfix)=($1, $2);

	my $isipv6;

	my @ips=_iptoipa($ip);

	$isipv6=shift @ips;

	croak "$pfix, as in '$cidr', does not make sense"
	    unless $pfix >= 0 && $pfix <= ($#ips+1) * 8 && $pfix =~ /^[0-9]+$/;

	my $i;

	for ($i=0; $i <= $#ips; $i++)
	{
	    last if $pfix - $i * 8 < 8;
	}

	my @msb=splice @ips, 0, $i;

	my $bitsleft= $pfix - $i * 8;

	if ($#ips < 0 || $bitsleft == 0)
	{
	    if ($pfix == 0 && $bitsleft == 0)
	    {
		foreach (0..255)
		{
		    my @n=($_);

		    if ($isipv6)
		    {
			_push_ipv6_octets(\@r, \@n);
		    }
		    else
		    {
			push @r, $n[0];
		    }
		}
	    }
	    elsif ($isipv6)
	    {
		_push_ipv6_octets(\@r, \@msb);
	    }
	    else
	    {
		push @r, join(".", @msb);
	    }
	    next;
	}

	my @rr=_cidr2range8(($ips[0], $bitsleft));

	while ($#rr >= 0)
	{
	    my $a=shift @rr;
	    my $b=shift @rr;

	    grep {
		if ($isipv6)
		{
		    push @msb, $_;
		    _push_ipv6_octets(\@r, \@msb);
		    pop @msb;
		}
		else
		{
		    push @r, join(".", (@msb, $_));
		}
	    } ($a .. $b);
	}
    }

    return @r;
}

sub _push_ipv6_octets {
    my $ary_ref=shift;
    my $octets=shift;

    if ( ($#{$octets} % 2) == 0)	# Odd number of octets
    {
	foreach (0 .. 255)
	{
	    push @$octets, $_;
	    _push_ipv6_octets($ary_ref, $octets);
	    pop @$octets;
	}
	return;
    }

    my $i;
    my $s="";

    for ($i=0; $i <= $#{$octets}; $i += 2)
    {
	$s .= ":" if $s ne "";
	$s .= sprintf("%02x%02x", $$octets[$i], $$octets[$i+1]);
    }
    push @$ary_ref, $s;
}

=pod

=head2 @cidr_list=Net::CIDR::cidradd($block, @cidr_list);

The cidradd() functions allows a CIDR list to be built one CIDR netblock
at a time, merging adjacent and overlapping ranges.
$block is a single netblock, expressed as either "start-finish", or
"address/prefix".
Example:

    @cidr_list=Net::CIDR::range2cidr("192.68.0.0-192.68.0.255");
    @cidr_list=Net::CIDR::cidradd("10.0.0.0/8", @cidr_list);
    @cidr_list=Net::CIDR::cidradd("192.68.1.0-192.68.1.255", @cidr_list);
				  
The result is a two-element array: ("10.0.0.0/8", "192.68.0.0/23").
IPv6 addresses are handled in an analogous fashion.

=cut

sub cidradd {
    my @cidr=@_;

    my $ip=shift @cidr;

    $ip="$ip-$ip" unless $ip =~ /[-\/]/;

    unshift @cidr, $ip;

    @cidr=cidr2range(@cidr);

    my @a;
    my @b;

    grep {
	croak "This doesn't look like start-end\n" unless /(.*)-(.*)/;
	push @a, $1;
	push @b, $2;
    } @cidr;

    my $lo=shift @a;
    my $hi=shift @b;

    my $i;

    for ($i=0; $i <= $#a; $i++)
    {
	last if _ipcmp($lo, $hi) > 0;

	next if _ipcmp($b[$i], $lo) < 0;
	next if _ipcmp($hi, $a[$i]) < 0;

	if (_ipcmp($a[$i],$lo) <= 0 && _ipcmp($hi, $b[$i]) <= 0)
	{
	    $lo=_add1($hi);
	    last;
	}

	if (_ipcmp($a[$i],$lo) <= 0)
	{
	    $lo=_add1($b[$i]);
	    next;
	}

	if (_ipcmp($hi, $b[$i]) <= 0)
	{
	    $hi=_sub1($a[$i]);
	    next;
	}

	$a[$i]=undef;
	$b[$i]=undef;
    }

    unless ((! defined $lo) || (! defined $hi) || _ipcmp($lo, $hi) > 0)
    {
	push @a, $lo;
	push @b, $hi;
    }

    @cidr=();

    @a=grep ( (defined $_), @a);
    @b=grep ( (defined $_), @b);

    for ($i=0; $i <= $#a; $i++)
    {
	push @cidr, "$a[$i]-$b[$i]";
    }

    @cidr=sort {
	$a =~ /(.*)-/;

	my $c=$1;

	$b =~ /(.*)-/;

	my $d=$1;

	my $e=_ipcmp($c, $d);
	return $e;
    } @cidr;

    $i=0;

    while ($i < $#cidr)
    {
	$cidr[$i] =~ /(.*)-(.*)/;

	my ($k, $l)=($1, $2);

	$cidr[$i+1] =~ /(.*)-(.*)/;

	my ($m, $n)=($1, $2);

	if (_ipcmp( _add1($l), $m) == 0)
	{
	    splice @cidr, $i, 2, "$k-$n";
	    next;
	}
	++$i;
    }

    return range2cidr(@cidr);
}


sub _add1 {
    my $n=shift;

    my $isipv6;

    ($isipv6, $n)=_ipv6to4($n);

    my @ip=split(/\./, $n);

    my $i=$#ip;

    while ($i >= 0)
    {
	last if ++$ip[$i] < 256;
	$ip[$i]=0;
	--$i;
    }

    return undef if $i < 0;

    $i=join(".", @ip);
    $i=_ipv4to6($i) if $isipv6;
    return $i;

}

sub _sub1 {
    my $n=shift;

    my $isipv6;

    ($isipv6, $n)=_ipv6to4($n);

    my @ip=split(/\./, $n);

    my $i=$#ip;

    while ($i >= 0)
    {
	last if --$ip[$i] >= 0;
	$ip[$i]=255;
	--$i;
    }

    return undef if $i < 0;

    $i=join(".", @ip);
    $i=_ipv4to6($i) if $isipv6;
    return $i;
}

=pod

=head2 $found=Net::CIDR::cidrlookup($ip, @cidr_list);

Search for $ip in @cidr_list.  $ip can be a single IP address, or a
netblock in CIDR or start-finish notation.
lookup() returns 1 if $ip overlaps any netblock in @cidr_list, 0 if not.

=cut

sub cidrlookup {
    my @cidr=@_;

    my $ip=shift @cidr;

    $ip="$ip-$ip" unless $ip =~ /[-\/]/;

    unshift @cidr, $ip;

    @cidr=cidr2range(@cidr);

    my @a;
    my @b;

    grep {
	croak "This doesn't look like start-end\n" unless /(.*)-(.*)/;
	push @a, $1;
	push @b, $2;
    } @cidr;

    my $lo=shift @a;
    my $hi=shift @b;

    my $i;

    for ($i=0; $i <= $#a; $i++)
    {
	next if _ipcmp($b[$i], $lo) < 0;
	next if _ipcmp($hi, $a[$i]) < 0;
	return 1;
    }

    return 0;
}

=pod

=head2 $ip=Net::CIDR::cidrvalidate($ip);

Validate whether $ip is a valid IPv4 or IPv6 address, or a CIDR.
Returns its argument or undef.
Spaces are removed, and IPv6 hexadecimal address are converted to lowercase.

$ip with less than four octets gets filled out with additional octets, and
the modified value gets returned. This turns "192.168/16" into a proper
"192.168.0.0/16".

If $ip contains a "/", it must be a valid CIDR, otherwise it must be a valid
IPv4 or an IPv6 address.

A technically invalid CIDR, such as "192.168.0.1/24" fails validation, returning
undef.

=cut

sub cidrvalidate {
    my $v=shift;

    $v =~ s/\s//g;

    $v=lc($v);

    my $suffix;

    ($v, $suffix)=($1, $2) if $v =~ m@(.*)/(.*)@;

    if (defined $suffix)
    {
	return undef unless $suffix =~ /^\d+$/ &&
	    ($suffix eq "0" || $suffix =~ /^[123456789]/);
    }

    if ($v =~ /^([0-9\.]+)$/ || $v =~ /^::ffff:([0-9\.]+)$/ ||
	$v =~ /^:([0-9\.]+)$/)
    {
	my $n=$1;

	return undef if $n =~ /^\./ || $n =~ /\.$/ || $n =~ /\.\./;

	my @o= split(/\./, $n);

	while ($#o < 3)
	{
	    push @o, "0";
	}

	$n=join(".", @o);

	return undef if $#o != 3;

	foreach (@o)
	{
	    return undef if /^0./;
	    return undef if $_ < 0 || $_ > 255;
	}

	if ($v =~ /^::ffff/)
	{
	    $suffix=128 unless defined $suffix;

	    return undef if $suffix < 128-32;

	    $suffix -= 128-32;
	}
	else
	{
	    $suffix=32 unless defined $suffix;
	}

	foreach (addr2cidr($n))
	{
	    return $_ if $_ eq "$n/$suffix";
	}
	return undef;
    }

    return undef unless $v =~ /^[0-9a-f:]+$/;

    return undef if $v =~ /:::/ || $v =~ /^:[^:]/ || $v =~ /[^:]:$/
	|| $v =~ /::.*::/;

    my @o=grep (/./, split(/:/, $v));

    return undef if ($#o >= 8 || ($#o<7 && $v !~ /::/));

    foreach (@o)
    {
	return undef if length ($_) > 4;
    }

    $suffix=128 unless defined $suffix;

    foreach (addr2cidr($v))
    {
	return $_ if $_ eq "$v/$suffix";
    }
    return undef;
}

=pod

=head1 BUGS

Garbage in, garbage out.
Always use cidrvalidate() before doing anything with untrusted input.
Otherwise,
"slightly" invalid input will work (extraneous whitespace
is generally OK),
but the functions will croak if you're totally off the wall.

=head1 AUTHOR

Sam Varshavchik <sam@email-scan.com>

With some contributions from David Cantrell <david@cantrell.org.uk>

=cut

__END__
	Global
`s`	Focus search bar
`?`	Bring up this help dialog
	GitHub
`g` `p`	Go to pull requests
`g` `i`	go to github issues (only if github is preferred repository)
	POD
`g` `a`	Go to author
`g` `c`	Go to changes
`g` `i`	Go to issues
`g` `d`	Go to dist
`g` `r`	Go to repository/SCM
`g` `s`	Go to source
`g` `b`	Go to file browse
	Search terms
module: (e.g. module:Plugin)
distribution: (e.g. distribution:Dancer auth)
author: (e.g. author:SONGMU Redis)
version: (e.g. version:1.00)