Update html-scanner to handle CDATA sections better. Closes #2970. [Jamis Buck]

git-svn-id: http://svn-commit.rubyonrails.org/rails/trunk@3153 5ecf4fe2-1ee6-0310-87b1-e25e094e27de

actionpack/CHANGELOG

@@ -1,5 +1,7 @@
 *SVN*
 
+* Update html-scanner to handle CDATA sections better. Closes #2970. [Jamis Buck]
+
 * Don't put flash in session if sessions are disabled.  [Jeremy Kemper]
 
 * Strip out trailing &_= for raw post bodies. Closes #2868. [Sam Stephenson]

actionpack/lib/action_controller/vendor/html-scanner/html/node.rb

@@ -150,6 +150,11 @@ module HTML #:nodoc:
             end
           end
 
+          if scanner.skip(/!\[CDATA\[/)
+            scanner.scan_until(/\]\]>/)
+            return CDATA.new(parent, line, pos, scanner.pre_match)
+          end
+          
           closing = ( scanner.scan(/\//) ? :close : nil )
           return Text.new(parent, line, pos, content) unless name = scanner.scan(/[\w:]+/)
           name.downcase!
@@ -257,6 +262,14 @@ module HTML #:nodoc:
     end
   end
   
+  # A CDATA node is simply a text node with a specialized way of displaying
+  # itself.
+  class CDATA < Text
+    def to_s
+      "<![CDATA[#{super}]>"
+    end
+  end
+
   # A Tag is any node that represents markup. It may be an opening tag, a
   # closing tag, or a self-closing tag. It has a name, and may have a hash of
   # attributes.

actionpack/lib/action_controller/vendor/html-scanner/html/tokenizer.rb

@@ -52,6 +52,9 @@ module HTML #:nodoc:
         if @scanner.scan(/!--/) # comment
           tag << @scanner.matched
           tag << (@scanner.scan_until(/--\s*>/) || @scanner.scan_until(/\Z/))
+        elsif @scanner.scan(/!\[CDATA\[/)
+          tag << @scanner.matched
+          tag << @scanner.scan_until(/\]\]>/)
         elsif @scanner.scan(/!/) # doctype
           tag << @scanner.matched
           tag << consume_quoted_regions