mlregex.rb

url = "http://www.google.com"
rex = %r{^(\w+)*://(\w+)?\.(\w+).?\.(\w+)}
url.match(rex)
url_data = {:host => $1, :hostname => $2, :domain => $3, :tld => $4}
p url_data # {:host=>"http", :hostname=>"www", :domain=>"google", :tld=>"com"} 

regex19.rb

url = "http://www.google.com"
rex = %r{
    ^
    (\w+)* # protocol
    ://
    (\w+)? # hostname
    \.
    (\w+)? # domain
    \.
    (\w+) # tld
}x
url.match(rex)
url_data = {:host => $1, :hostname => $2, :domain => $3, :tld => $4}
p url_data # {:host=>"http", :hostname=>"www", :domain=>"google", :tld=>"com"} 

regex192.rb

url = "http://www.google.com"
rex = %r{
  ^                 #starts with
  (?<protocol>\w+)*   #either a protocol or not
  ://
  (?<hostname>\w+)?
  \.
  (?<domain>\w+)?
  \.
  (?<tld>\w+)?
}x
result = url.match(rex)
p result # #<MatchData "http://www.google.com" protocol:"http" hostname:"www" domain:"google" tld:"com">

regex192os.rb

require 'ostruct'
url = "http://www.google.com"
rex = %r{
  ^                 #starts with
  (?<protocol>\w+)*   #either a protocol or not
  ://
  (?<hostname>\w+)?
  \.
  (?<domain>\w+)?
  \.
  (?<tld>\w+)?
}x
result = url.match(rex)
class MatchData
  def to_hash
    Hash[*names.zip(captures).flatten]    
  end
end
URL = Class.new(OpenStruct)
url = URL.new(result.to_hash)
p url #  #<URL protocol="http", hostname="www", domain="google", tld="com">