regular expressions are usually ASCII-encoded, so force_encoding the content of a Node to the encoding of the regular expression is wrong.

2 files changed, 1 insertions, 2 deletions

diff --git a/actionpack/lib/action_controller/vendor/html-scanner/html/tokenizer.rb b/actionpack/lib/action_controller/vendor/html-scanner/html/tokenizer.rb
index 602411ed37..064ff3724d 100644
--- a/actionpack/lib/action_controller/vendor/html-scanner/html/tokenizer.rb
+++ b/actionpack/lib/action_controller/vendor/html-scanner/html/tokenizer.rb
@@ -23,6 +23,7 @@ module HTML #:nodoc:
     
     # Create a new Tokenizer for the given text.
     def initialize(text)
+      text.encode! if text.encoding_aware?
       @scanner = StringScanner.new(text)
       @position = 0
       @line = 0
diff --git a/actionpack/lib/action_dispatch/testing/assertions/selector.rb b/actionpack/lib/action_dispatch/testing/assertions/selector.rb
index 9deabf5b3c..0e82b41590 100644
--- a/actionpack/lib/action_dispatch/testing/assertions/selector.rb
+++ b/actionpack/lib/action_dispatch/testing/assertions/selector.rb
@@ -267,14 +267,12 @@ module ActionDispatch
         if match_with = equals[:text]
           matches.delete_if do |match|
             text = ""
-            text.force_encoding(match_with.encoding) if text.respond_to?(:force_encoding)
             stack = match.children.reverse
             while node = stack.pop
               if node.tag?
                 stack.concat node.children.reverse
               else
                 content = node.content
-                content.force_encoding(match_with.encoding) if content.respond_to?(:force_encoding)
                 text << content
               end
             end