package Tantrix::SLDfind;

use warnings;
use strict;

# Country code TLDs. http://www.iana.org/cctld/cctld-whois.htm
my $cctld = '[a-z][a-z]'; # close enough :)

# Three letter Generic TLDs. http://www.iana.org/gtld/gtld.htm
# Spam is unlikely from TLDs reserved for
# governments or treaties between governments.
# my $tld3 = '(?:biz|com|net|org|pro|gov|edu|mil|int)';
my $gtld = '(?:biz|com|net|org)';

# Second level domains. Prefer to have (?:co|ne|...) etc but we can catch
# additional oddities like 'uk.com' this way.
my $sld2 = '[a-z][a-z]';
my $sld3 = '[a-z][a-z][a-z]';

# Four letter generic TLDs. http://www.iana.org/gtld/gtld.htm
# Until spam appears on any besides .info, ignore them.
# my $tld4 = '(?:info|aero|coop|name)';
my $gtld4 = 'info';

# Any char that can appear in a domain
my $dc   = '[-a-z0-9]';
my $dc4p = "$dc\{4,}";	# Four or more of 'em

# No more domain char-like characters
my $end  = "(?![-a-z0-9.@])";

# Observe that spamvertised domains are unlikely to be valuable three
# letter domains. (Even four, frankly, is unlikely.)
our $domain_find_re = qr/
	(?:
		(?:$dc4p \. (?:$sld3\.$cctld	# com.mx
	               |   $sld2\.$gtld 	# uk.com
	               |   $sld2\.$cctld	# co.nz
	               |          $gtld 	# net
	               |          $cctld)	# de
	    )
	|	(?:$dc+  \. $gtld4)          	# info
	) $end
/xi;

1;