# Script to spider a web application at random
# 
# You'll need rfuzz and hpricot to run this script:
# gem install rfuzz
# gem install hpricot
require 'rubygems'
require 'hpricot'
require 'rfuzz/client'

cl = RFuzz::HttpClient.new('localhost', 4000)

# First path to visit
paths = ['/people']

# Paths already visited
visited_paths = []

while !paths.empty?
  # Get the next path to visit
  path = paths.pop
  
  # Ignore it if already visited
  if visited_paths.member?(path)
    puts "Already visited #{path} - ignoring"
  else
    visited_paths.push path

    puts "Getting links from page at #{path}"    
    body = cl.get(path).http_body
    doc = Hpricot(body)
    
    # Get all links out of the page
    links = doc.search("//a")
    links.each do |l|
      # Add the path for each URL into array of paths to visit
      found_path = l.attributes['href']
      
      # Ignore mailto and empty links
      unless /^mailto:/ =~ found_path or found_path.empty?
        puts "Found link #{found_path}"
        paths.push(found_path)
      end
    end
    
    # Visit current path 10-100 times with get requests
    num_gets = 10 + (rand(9) * 10)
    puts "Visiting #{path} #{num_gets} times"
    (1..num_gets).each do |i|
      cl.get(path)
    end
  end
end