Safe parsing of javascript links

google.rb

@@ -1,40 +1,45 @@
+# frozen_string_literal: true
+
 require 'mechanize'
 agent = Mechanize.new
 
-def check_page_for_links(indexed_pages=[],address,agent)
-  begin
-    page = agent.get(address);
-    page.links().each do | link |
-      link = link.resolved_uri;
-      link.fragment = nil;
-      link.query = nil;
-      if !(indexed_pages.include?(link))
-        if link.host == address.host
-          #p link
-          indexed_pages.push(link)
-        end
+# Function to check for links
+def check_for_links(address, agent, indexed_pages = [])
+  # Check all links on the page
+  page = agent.get(address)
+  if page.is_a?(Mechanize::Page)
+    page.links.each do |link|
+      # Get a complete resolved URI for the link
+      next if link.href == 'javascript:'
+      link = link.resolved_uri
+      # Remove the fragment so there's less overlap
+      link.fragment = nil
+      # Clear query (TODO: Stash away later)
+      link.query = nil
+      # Check if we've seen the link before
+      # Check if we're leaving the host
+      if !indexed_pages.include?(link) && link.host == address.host
+        # Save link to array
+        indexed_pages.push(link)
       end
     end
-    return indexed_pages;
-  rescue Mechanize::ResponseCodeError => e
-    p e.response_code
-  rescue 
+    # Return all functions
+    indexed_pages
+  end
+rescue Mechanize::ResponseCodeError => e
+  p e.response_code
+end
+
+addresses = check_for_links(URI.parse('http://git.blizzard.systems'), agent)
+old_addresses = []
+until addresses - old_addresses == []
+  old_addresses = Array.new(addresses)
+  addresses.each do |address|
+    appending_addrs = check_for_links(address, agent, addresses)
+    addresses.union(appending_addrs) unless appending_addrs.nil?
   end
 end
 
-addresses = check_page_for_links([],URI::parse('http://git.blizzard.systems'),agent);
-old_addresses = Array.new();
-#p addresses;
-until addresses - old_addresses == [] do
-  old_addresses = Array.new(addresses);
-  addresses.each do | address |
-    appending_addrs = check_page_for_links(addresses,address,agent);
-    if !appending_addrs.nil? 
-      addresses.union(appending_addrs);
-    end
-  end
-end
-
-p addresses;
-puts "********************************************";
-p old_addresses;
+p addresses
+puts '********************************************'
+p old_addresses

lib/fuzzer.rb

@@ -23,22 +23,26 @@ end
 def check_for_links(address, agent, indexed_pages = [])
   # Check all links on the page
   page = agent.get(address)
-  page.links.each do |link|
-    # Get a complete resolved URI for the link
-    link = link.resolved_uri
-    # Remove the fragment so there's less overlap
-    link.fragment = nil
-    # Clear query (TODO: Stash away later)
-    link.query = nil
-    # Check if we've seen the link before
-    # Check if we're leaving the host
-    if !indexed_pages.include?(link) && link.host == address.host
-      # Save link to array
-      indexed_pages.push(link)
+  if page.is_a?(Mechanize::Page)
+    page.links.each do |link|
+      # Get a complete resolved URI for the link
+      next if link.href == 'javascript:'
+
+      link = link.resolved_uri
+      # Remove the fragment so there's less overlap
+      link.fragment = nil
+      # Clear query (TODO: Stash away later)
+      link.query = nil
+      # Check if we've seen the link before
+      # Check if we're leaving the host
+      if !indexed_pages.include?(link) && link.host == address.host
+        # Save link to array
+        indexed_pages.push(link)
+      end
     end
+    # Return all functions
+    indexed_pages
   end
-  # Return all functions
-  indexed_pages
 rescue Mechanize::ResponseCodeError => e
   p e.response_code
 end