## mark a section (Learn more)

#!/usr/bin/env ruby -wKU EMAIL_PATTERN = /\b[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}\b/i module HTML class Node include Enumerable def initialize @children = [ ] end def <<(child) @children << child end def self.prune=(flag) @@prune = flag end def each(&block) @@prune = false if self.class == Node block.call(self) unless self.class == Node @children.any? { |e| e.each(&block) } unless @@prune res = @@prune @@prune = false res end end class Tag < Node attr_writer :close_tag # TODO we should raise if an unexpected end tag is provided attr_reader :args def initialize(tag) super() @open_tag, @close_tag = tag, '' args = tag.gsub(/^<[a-z0-9]+\s*|\s*>$/i, '') @args = Hash[*args.scan(/([a-z0-9]+)\s*=\s*("(?:[^<"]+|<\?.*?\?>)*?"|'(?:[^<']+|<\?.*?\?>)*?')/m).flatten] end def each(&block) unless super(&block) block.call(@close_tag) unless @@prune end end def tag $&.downcase if @open_tag =~ /[a-z0-9]+/i end def to_s @open_tag end end class Text < Node def initialize(text) super() @text = text end def to_s @text end end module_function # Input: an IO object that holds an HTML page optionally with <% script %> <?php tags ?> (this is why an off-the-shelves parser was no good) # Output: a Node object implementing the Enumerable interface (this node is the root of the tree representing the parsed page) # the parser is regexp-based so not overly robust def parse(io) # TODO handle XHTML # TODO support <div arg="<% … %>"> (we presently only support <? … ?>) pattern = / ( <(?:br|hr|img|meta|link|input|base|area|col|frame|param)\b[^>]*> ) # single tag | ( < [a-z0-9]+ (?: [^<>]+ | <\?.*?\?> )* > ) # open tag | ( < \/ [a-z0-9]+ [^>]* > ) # close tag | ( [^<]+ ) # text | ( <!DOCTYPE\b [^>]* > | <!--.*?--> | <\?.*?\?> | <%.*?%> ) # stuff to preserve /xmi stack = [ Node.new ] io.read.scan(pattern) do |m| if single_tag = m[0] stack[-1] << Tag.new(single_tag) elsif start_tag = m[1] stack.push(Tag.new(start_tag)) stack[-2] << stack[-1] elsif close_tag = m[2] raise "Close tag found but no tags are open" if stack.size == 0 stack[-1].close_tag = close_tag stack.pop elsif text = m[3] stack[-1] << Text.new(text) elsif preserve = m[4] stack[-1] << Text.new(preserve) else raise "Unhandled construct: " + m[5].split("\n").first end end raise "Reaching end of document with unclosed tags (#{stack[1..-1].join(", ")})" unless stack.size == 1 stack.first end end def e_js(str) str.gsub(/(?=[\\"])/, '\\').gsub(/\n/, '\n').gsub(/[@.\/]/) { |ch| sprintf('\\%03o', ch[0]) } end def rot_13(str) str.tr('A-Za-z', 'N-ZA-Mn-za-m') end def obfuscate(text) %{<script type="text/javascript">document.write(} + %{"#{e_js(rot_13(text))}".replace(/[a-zA-Z]/g, function(c){return String.fromCharCode((c<="Z"?90:122)>=(c=c.charCodeAt(0)+13)?c:c-26);}));} + %{</script><noscript><em>(sorry, we are protecting ourself against spam so JavaScript is required to show the email address)</em></noscript>} end HTML.parse(STDIN).map do |e| # to_s will give only the text of the current object, i.e. <tag> for <tag>foo</tag> # to_a will return the sub-tree as an array, i.e. [ <tag>, foo, </tag> ] if e.to_s =~ EMAIL_PATTERN STDOUT << obfuscate(e.to_a.join) HTML::Node.prune = true # skip descending into current sub-tree else STDOUT << e.to_s end end

This paste will be private.