From 244f96f089e3aa7dc19b0382314d3c5cd6ef1863 Mon Sep 17 00:00:00 2001
From: Mike Dalessio <mike@csa.net>
Date: Tue, 9 Mar 2010 01:29:36 -0500
Subject: [PATCH] allowing #text to return unescaped html entities if the user
 very carefully asks for it.

---
 lib/loofah/html/document.rb          | 24 +++++++++++++++---
 lib/loofah/html/document_fragment.rb | 24 +++++++++++++++---
 test/integration/test_scrubbers.rb   | 37 ++++++++++++++++++++++++++++
 3 files changed, 77 insertions(+), 8 deletions(-)

diff --git a/lib/loofah/html/document.rb b/lib/loofah/html/document.rb
index d7b4fc76..e70381a9 100644
--- a/lib/loofah/html/document.rb
+++ b/lib/loofah/html/document.rb
@@ -19,8 +19,24 @@ class Document < Nokogiri::HTML::Document
       #    Loofah.document("<h1>Title</h1><div>Content</div>").text
       #    # => "TitleContent"
       #
-      def text
-        encode_special_chars xpath("/html/body").inner_text
+      #  By default, the returned text will have HTML entities
+      #  escaped. If you want unescaped entities, and you understand
+      #  that the result is unsafe to render in a browser, then you
+      #  can pass an argument as shown:
+      #
+      #    frag = Loofah.fragment("&lt;script&gt;alert('EVIL');&lt;/script&gt;")
+      #    # ok for browser:
+      #    frag.text                                 # => "&lt;script&gt;alert('EVIL');&lt;/script&gt;"
+      #    # decidedly not ok for browser:
+      #    frag.text(:encode_special_chars => false) # => "<script>alert('EVIL');</script>"
+      #
+      def text(options={})
+        result = xpath("/html/body").inner_text
+        if options[:encode_special_chars] == false
+          result # possibly dangerous if rendered in a browser
+        else
+          encode_special_chars result
+        end
       end
       alias :inner_text :text
       alias :to_str     :text
@@ -35,8 +51,8 @@ def text
       #    Loofah.document("<h1>Title</h1><div>Content</div>").to_text
       #    # => "\nTitle\n\nContent\n"
       #
-      def to_text
-        Loofah::Helpers.remove_extraneous_whitespace self.dup.scrub!(:newline_block_elements).text
+      def to_text(options={})
+        Loofah::Helpers.remove_extraneous_whitespace self.dup.scrub!(:newline_block_elements).text(options)
       end
     end
   end
diff --git a/lib/loofah/html/document_fragment.rb b/lib/loofah/html/document_fragment.rb
index 87fac7c9..12562455 100644
--- a/lib/loofah/html/document_fragment.rb
+++ b/lib/loofah/html/document_fragment.rb
@@ -35,8 +35,24 @@ def to_s
       #    Loofah.fragment("<h1>Title</h1><div>Content</div>").text
       #    # => "TitleContent"
       #
-      def text
-        encode_special_chars serialize_roots.children.inner_text
+      #  By default, the returned text will have HTML entities
+      #  escaped. If you want unescaped entities, and you understand
+      #  that the result is unsafe to render in a browser, then you
+      #  can pass an argument as shown:
+      #
+      #    frag = Loofah.fragment("&lt;script&gt;alert('EVIL');&lt;/script&gt;")
+      #    # ok for browser:
+      #    frag.text                                 # => "&lt;script&gt;alert('EVIL');&lt;/script&gt;"
+      #    # decidedly not ok for browser:
+      #    frag.text(:encode_special_chars => false) # => "<script>alert('EVIL');</script>"
+      #
+      def text(options={})
+        result = serialize_roots.children.inner_text
+        if options[:encode_special_chars] == false
+          result # possibly dangerous if rendered in a browser
+        else
+          encode_special_chars result
+        end
       end
       alias :inner_text :text
       alias :to_str     :text
@@ -51,8 +67,8 @@ def text
       #    Loofah.fragment("<h1>Title</h1><div>Content</div>").to_text
       #    # => "\nTitle\n\nContent\n"
       #
-      def to_text
-        Loofah::Helpers.remove_extraneous_whitespace self.dup.scrub!(:newline_block_elements).text
+      def to_text(options={})
+        Loofah::Helpers.remove_extraneous_whitespace self.dup.scrub!(:newline_block_elements).text(options)
       end
 
       private
diff --git a/test/integration/test_scrubbers.rb b/test/integration/test_scrubbers.rb
index 9c68b48f..1278f15d 100644
--- a/test/integration/test_scrubbers.rb
+++ b/test/integration/test_scrubbers.rb
@@ -18,6 +18,7 @@ class TestScrubbers < Test::Unit::TestCase
 
   ENTITY_HACK_ATTACK            = "<div><div>Hack attack!</div><div>&lt;script&gt;alert('evil')&lt;/script&gt;</div></div>"
   ENTITY_HACK_ATTACK_TEXT_SCRUB = "Hack attack!&lt;script&gt;alert('evil')&lt;/script&gt;"
+  ENTITY_HACK_ATTACK_TEXT_SCRUB_UNESC = "Hack attack!<script>alert('evil')</script>"
 
   context "Document" do
     context "#scrub!" do
@@ -89,6 +90,24 @@ class TestScrubbers < Test::Unit::TestCase
 
         assert_equal ENTITY_HACK_ATTACK_TEXT_SCRUB, result
       end
+
+      context "with encode_special_chars => false" do
+        should "leave behind only inner text with html entities unescaped" do
+          doc = Loofah::HTML::Document.parse "<html><body>#{ENTITY_HACK_ATTACK}</body></html>"
+          result = doc.text(:encode_special_chars => false)
+
+          assert_equal ENTITY_HACK_ATTACK_TEXT_SCRUB_UNESC, result
+        end
+      end
+
+      context "with encode_special_chars => true" do
+        should "leave behind only inner text with html entities still escaped" do
+          doc = Loofah::HTML::Document.parse "<html><body>#{ENTITY_HACK_ATTACK}</body></html>"
+          result = doc.text(:encode_special_chars => true)
+
+          assert_equal ENTITY_HACK_ATTACK_TEXT_SCRUB, result
+        end
+      end
     end
 
     context "#to_s" do
@@ -239,6 +258,24 @@ class TestScrubbers < Test::Unit::TestCase
 
         assert_equal ENTITY_HACK_ATTACK_TEXT_SCRUB, result
       end
+
+      context "with encode_special_chars => false" do
+        should "leave behind only inner text with html entities unescaped" do
+          doc = Loofah::HTML::DocumentFragment.parse "<div>#{ENTITY_HACK_ATTACK}</div>"
+          result = doc.text(:encode_special_chars => false)
+
+          assert_equal ENTITY_HACK_ATTACK_TEXT_SCRUB_UNESC, result
+        end
+      end
+
+      context "with encode_special_chars => true" do
+        should "leave behind only inner text with html entities still escaped" do
+          doc = Loofah::HTML::DocumentFragment.parse "<div>#{ENTITY_HACK_ATTACK}</div>"
+          result = doc.text(:encode_special_chars => true)
+
+          assert_equal ENTITY_HACK_ATTACK_TEXT_SCRUB, result
+        end
+      end
     end
 
     context "#to_s" do