rails-html-sanitizer 1.4.3 → 1.4.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 2f00d9f256478eb753c8d211c3b25efa4204bbdbc9c5abf0415413c811a2e404
-  data.tar.gz: 65d3871aa798dfbbfb1138b666d475b590e347cdb66614d6d39b72ad3531c742
+  metadata.gz: a74021096590326ee357971bec71d2c4507a95cdaf05c8e21d383ce18fee18d3
+  data.tar.gz: faad0d5f268dad601b633b03912e353fcc2d760fceb253d9cde2064b010b997a
 SHA512:
-  metadata.gz: e6e31eaa72b1a2e8356aae50600ac784f85a80828cbc49ce8061384ecd3f21a1d8eaee69845dc08537c5102728c3cc41a72cb3ed8b9789c4921038398afa61e2
-  data.tar.gz: 6b14a49842eaf4c3e0fbae5acd28fdf32a5deb6cd42f769aada848226847180c4d3a67a9dcbc439e1a4855699b0ea694cb4c7b6ee173391ac841bd334ae44b6f
+  metadata.gz: e7f01438708076a283326c78b052ba954a42de4134d8d1d7e7c336c82ecd04c661f75dad3a0f9b1ffebe278f76ef229c98a3f2568801f82d94c94a50f399a2ef
+  data.tar.gz: 4f44c0e92eb9e565611772ba28d426025621c0517c4217004c3409192991a17498dd38165a6c55561a5347d2fcdf34f51b24101ad6de525604e35785e89efbc0

data/CHANGELOG.md

@@ -1,3 +1,38 @@
+## 1.4.4 / 2022-12-13
+
+* Address inefficient regular expression complexity with certain configurations of Rails::Html::Sanitizer.
+
+  Fixes CVE-2022-23517. See
+  [GHSA-5x79-w82f-gw8w](https://github.com/rails/rails-html-sanitizer/security/advisories/GHSA-5x79-w82f-gw8w)
+  for more information.
+
+  *Mike Dalessio*
+
+* Address improper sanitization of data URIs.
+
+  Fixes CVE-2022-23518 and #135. See
+  [GHSA-mcvf-2q2m-x72m](https://github.com/rails/rails-html-sanitizer/security/advisories/GHSA-mcvf-2q2m-x72m)
+  for more information.
+
+  *Mike Dalessio*
+
+* Address possible XSS vulnerability with certain configurations of Rails::Html::Sanitizer.
+
+  Fixes CVE-2022-23520. See
+  [GHSA-rrfc-7g8p-99q8](https://github.com/rails/rails-html-sanitizer/security/advisories/GHSA-rrfc-7g8p-99q8)
+  for more information.
+
+  *Mike Dalessio*
+
+* Address possible XSS vulnerability with certain configurations of Rails::Html::Sanitizer.
+
+  Fixes CVE-2022-23519. See
+  [GHSA-9h9g-93gc-623h](https://github.com/rails/rails-html-sanitizer/security/advisories/GHSA-9h9g-93gc-623h)
+  for more information.
+
+  *Mike Dalessio*
+
+
 ## 1.4.3 / 2022-06-09
 
 * Address a possible XSS vulnerability with certain configurations of Rails::Html::Sanitizer.

data/lib/rails/html/sanitizer/version.rb

@@ -1,7 +1,7 @@
 module Rails
   module Html
     class Sanitizer
-      VERSION = "1.4.3"
+      VERSION = "1.4.4"
     end
   end
 end

data/lib/rails/html/sanitizer.rb

@@ -141,25 +141,8 @@ module Rails
 
       private
 
-      def loofah_using_html5?
-        # future-proofing, see https://github.com/flavorjones/loofah/pull/239
-        Loofah.respond_to?(:html5_mode?) && Loofah.html5_mode?
-      end
-
-      def remove_safelist_tag_combinations(tags)
-        if !loofah_using_html5? && tags.include?("select") && tags.include?("style")
-          warn("WARNING: #{self.class}: removing 'style' from safelist, should not be combined with 'select'")
-          tags.delete("style")
-        end
-        tags
-      end
-
       def allowed_tags(options)
-        if options[:tags]
-          remove_safelist_tag_combinations(options[:tags])
-        else
-          self.class.allowed_tags
-        end
+        options[:tags] || self.class.allowed_tags
       end
 
       def allowed_attributes(options)

data/lib/rails/html/scrubbers.rb

@@ -61,9 +61,9 @@ module Rails
       end
 
       def scrub(node)
-        if node.cdata?
-          text = node.document.create_text_node node.text
-          node.replace text
+        if Loofah::HTML5::Scrub.cdata_needs_escaping?(node)
+          replacement = Loofah::HTML5::Scrub.cdata_escape(node)
+          node.replace(replacement)
           return CONTINUE
         end
         return CONTINUE if skip_node?(node)
@@ -139,15 +139,13 @@ module Rails
                     end
 
         if Loofah::HTML5::SafeList::ATTR_VAL_IS_URI.include?(attr_name)
-          # this block lifted nearly verbatim from HTML5 sanitization
-          val_unescaped = CGI.unescapeHTML(attr_node.value).gsub(Loofah::HTML5::Scrub::CONTROL_CHARACTERS,'').downcase
-          if val_unescaped =~ /^[a-z0-9][-+.a-z0-9]*:/ && ! Loofah::HTML5::SafeList::ALLOWED_PROTOCOLS.include?(val_unescaped.split(Loofah::HTML5::SafeList::PROTOCOL_SEPARATOR)[0])
-            attr_node.remove
-          end
+          return if Loofah::HTML5::Scrub.scrub_uri_attribute(attr_node)
         end
+
         if Loofah::HTML5::SafeList::SVG_ATTR_VAL_ALLOWS_REF.include?(attr_name)
-          attr_node.value = attr_node.value.gsub(/url\s*\(\s*[^#\s][^)]+?\)/m, ' ') if attr_node.value
+          Loofah::HTML5::Scrub.scrub_attribute_that_allows_local_ref(attr_node)
         end
+
         if Loofah::HTML5::SafeList::SVG_ALLOW_LOCAL_HREF.include?(node.name) && attr_name == 'xlink:href' && attr_node.value =~ /^\s*[^#\s].*/m
           attr_node.remove
         end

data/test/sanitizer_test.rb

@@ -54,7 +54,7 @@ class SanitizersTest < Minitest::Test
 
   def test_strip_tags_with_quote
     input = '<" <img src="trollface.gif" onload="alert(1)"> hi'
-    expected = libxml_2_9_14_recovery? ? %{&lt;"  hi} : %{ hi}
+    expected = libxml_2_9_14_recovery_lt? ? %{&lt;"  hi} : %{ hi}
     assert_equal(expected, full_sanitize(input))
   end
 
@@ -77,19 +77,19 @@ class SanitizersTest < Minitest::Test
 
   def test_remove_unclosed_tags
     input = "This is <-- not\n a comment here."
-    expected = libxml_2_9_14_recovery? ? %{This is &lt;-- not\n a comment here.} : %{This is }
+    expected = libxml_2_9_14_recovery_lt? ? %{This is &lt;-- not\n a comment here.} : %{This is }
     assert_equal(expected, full_sanitize(input))
   end
 
   def test_strip_cdata
     input = "This has a <![CDATA[<section>]]> here."
-    expected = libxml_2_9_14_recovery? ? %{This has a &lt;![CDATA[]]&gt; here.} : %{This has a ]]&gt; here.}
+    expected = libxml_2_9_14_recovery_lt_bang? ? %{This has a &lt;![CDATA[]]&gt; here.} : %{This has a ]]&gt; here.}
     assert_equal(expected, full_sanitize(input))
   end
 
   def test_strip_unclosed_cdata
     input = "This has an unclosed <![CDATA[<section>]] here..."
-    expected = libxml_2_9_14_recovery? ? %{This has an unclosed &lt;![CDATA[]] here...} : %{This has an unclosed ]] here...}
+    expected = libxml_2_9_14_recovery_lt_bang? ? %{This has an unclosed &lt;![CDATA[]] here...} : %{This has an unclosed ]] here...}
     assert_equal(expected, full_sanitize(input))
   end
 
@@ -458,13 +458,13 @@ class SanitizersTest < Minitest::Test
 
   def test_should_sanitize_cdata_section
     input = "<![CDATA[<span>section</span>]]>"
-    expected = libxml_2_9_14_recovery? ? %{&lt;![CDATA[<span>section</span>]]&gt;} : %{section]]&gt;}
+    expected = libxml_2_9_14_recovery_lt_bang? ? %{&lt;![CDATA[<span>section</span>]]&gt;} : %{section]]&gt;}
     assert_sanitized(input, expected)
   end
 
   def test_should_sanitize_unterminated_cdata_section
     input = "<![CDATA[<span>neverending..."
-    expected = libxml_2_9_14_recovery? ? %{&lt;![CDATA[<span>neverending...</span>} : %{neverending...}
+    expected = libxml_2_9_14_recovery_lt_bang? ? %{&lt;![CDATA[<span>neverending...</span>} : %{neverending...}
     assert_sanitized(input, expected)
   end
 
@@ -581,23 +581,124 @@ class SanitizersTest < Minitest::Test
     assert_equal("<div>text</div><b>text</b>", safe_list_sanitize("<div>text</div><!-- comment --><b>text</b>"))
   end
 
-  def test_disallow_the_dangerous_safelist_combination_of_select_and_style
-    input = "<select><style><script>alert(1)</script></style></select>"
-    tags = ["select", "style"]
-    warning = /WARNING: Rails::Html::SafeListSanitizer: removing 'style' from safelist/
-    sanitized = nil
-    invocation = Proc.new { sanitized = safe_list_sanitize(input, tags: tags) }
+  %w[text/plain text/css image/png image/gif image/jpeg].each do |mediatype|
+    define_method "test_mediatype_#{mediatype}_allowed" do
+      input = %Q(<img src="data:#{mediatype};base64,PHNjcmlwdD5hbGVydCgnWFNTJyk8L3NjcmlwdD4=">)
+      expected = input
+      actual = safe_list_sanitize(input)
+      assert_equal(expected, actual)
 
-    if html5_mode?
-      # if Loofah is using an HTML5 parser,
-      #   then "style" should be removed by the parser as an invalid child of "select"
-      assert_silent(&invocation)
-    else
-      # if Loofah is using an HTML4 parser,
-      #   then SafeListSanitizer should remove "style" from the safelist
-      assert_output(nil, warning, &invocation)
+      input = %Q(<img src="DATA:#{mediatype};base64,PHNjcmlwdD5hbGVydCgnWFNTJyk8L3NjcmlwdD4=">)
+      expected = input
+      actual = safe_list_sanitize(input)
+      assert_equal(expected, actual)
+    end
+  end
+
+  def test_mediatype_text_html_disallowed
+    input = %q(<img src="data:text/html;base64,PHNjcmlwdD5hbGVydCgnWFNTJyk8L3NjcmlwdD4=">)
+    expected = %q(<img>)
+    actual = safe_list_sanitize(input)
+    assert_equal(expected, actual)
+
+    input = %q(<img src="DATA:text/html;base64,PHNjcmlwdD5hbGVydCgnWFNTJyk8L3NjcmlwdD4=">)
+    expected = %q(<img>)
+    actual = safe_list_sanitize(input)
+    assert_equal(expected, actual)
+  end
+
+  def test_mediatype_image_svg_xml_disallowed
+    input = %q(<img src="data:image/svg+xml;base64,PHNjcmlwdD5hbGVydCgnWFNTJyk8L3NjcmlwdD4=">)
+    expected = %q(<img>)
+    actual = safe_list_sanitize(input)
+    assert_equal(expected, actual)
+
+    input = %q(<img src="DATA:image/svg+xml;base64,PHNjcmlwdD5hbGVydCgnWFNTJyk8L3NjcmlwdD4=">)
+    expected = %q(<img>)
+    actual = safe_list_sanitize(input)
+    assert_equal(expected, actual)
+  end
+
+  def test_mediatype_other_disallowed
+    input = %q(<a href="data:foo;base64,PHNjcmlwdD5hbGVydCgnWFNTJyk8L3NjcmlwdD4=">foo</a>)
+    expected = %q(<a>foo</a>)
+    actual = safe_list_sanitize(input)
+    assert_equal(expected, actual)
+
+    input = %q(<a href="DATA:foo;base64,PHNjcmlwdD5hbGVydCgnWFNTJyk8L3NjcmlwdD4=">foo</a>)
+    expected = %q(<a>foo</a>)
+    actual = safe_list_sanitize(input)
+    assert_equal(expected, actual)
+  end
+
+  def test_scrubbing_svg_attr_values_that_allow_ref
+    input = %Q(<div fill="yellow url(http://bad.com/) #fff">hey</div>)
+    expected = %Q(<div fill="yellow #fff">hey</div>)
+    actual = scope_allowed_attributes %w(fill) do
+      safe_list_sanitize(input)
     end
-    refute_includes(sanitized, "style")
+
+    assert_equal(expected, actual)
+  end
+
+  def test_style_with_css_payload
+    input, tags = "<style>div > span { background: \"red\"; }</style>", ["style"]
+    expected = "<style>div &gt; span { background: \"red\"; }</style>"
+    actual = safe_list_sanitize(input, tags: tags)
+
+    assert_equal(expected, actual)
+  end
+
+  def test_combination_of_select_and_style_with_css_payload
+    input, tags = "<select><style>div > span { background: \"red\"; }</style></select>", ["select", "style"]
+    expected = "<select><style>div &gt; span { background: \"red\"; }</style></select>"
+    actual = safe_list_sanitize(input, tags: tags)
+
+    assert_equal(expected, actual)
+  end
+
+  def test_combination_of_select_and_style_with_script_payload
+    input, tags = "<select><style><script>alert(1)</script></style></select>", ["select", "style"]
+    expected = "<select><style>&lt;script&gt;alert(1)&lt;/script&gt;</style></select>"
+    actual = safe_list_sanitize(input, tags: tags)
+
+    assert_equal(expected, actual)
+  end
+
+  def test_combination_of_svg_and_style_with_script_payload
+    input, tags = "<svg><style><script>alert(1)</script></style></svg>", ["svg", "style"]
+    expected = "<svg><style>&lt;script&gt;alert(1)&lt;/script&gt;</style></svg>"
+    actual = safe_list_sanitize(input, tags: tags)
+
+    assert_equal(expected, actual)
+  end
+
+  def test_combination_of_math_and_style_with_img_payload
+    input, tags = "<math><style><img src=x onerror=alert(1)></style></math>", ["math", "style"]
+    expected = "<math><style>&lt;img src=x onerror=alert(1)&gt;</style></math>"
+    actual = safe_list_sanitize(input, tags: tags)
+
+    assert_equal(expected, actual)
+
+    input, tags = "<math><style><img src=x onerror=alert(1)></style></math>", ["math", "style", "img"]
+    expected = "<math><style>&lt;img src=x onerror=alert(1)&gt;</style></math>"
+    actual = safe_list_sanitize(input, tags: tags)
+
+    assert_equal(expected, actual)
+  end
+
+  def test_combination_of_svg_and_style_with_img_payload
+    input, tags = "<svg><style><img src=x onerror=alert(1)></style></svg>", ["svg", "style"]
+    expected = "<svg><style>&lt;img src=x onerror=alert(1)&gt;</style></svg>"
+    actual = safe_list_sanitize(input, tags: tags)
+
+    assert_equal(expected, actual)
+
+    input, tags = "<svg><style><img src=x onerror=alert(1)></style></svg>", ["svg", "style", "img"]
+    expected = "<svg><style>&lt;img src=x onerror=alert(1)&gt;</style></svg>"
+    actual = safe_list_sanitize(input, tags: tags)
+
+    assert_equal(expected, actual)
   end
 
 protected
@@ -657,11 +758,14 @@ protected
     end.join
   end
 
-  def libxml_2_9_14_recovery?
+  def libxml_2_9_14_recovery_lt?
+    # changed in 2.9.14, see https://github.com/sparklemotion/nokogiri/releases/tag/v1.13.5
     Nokogiri.method(:uses_libxml?).arity == -1 && Nokogiri.uses_libxml?(">= 2.9.14")
   end
 
-  def html5_mode?
-    ::Loofah.respond_to?(:html5_mode?) && ::Loofah.html5_mode?
+  def libxml_2_9_14_recovery_lt_bang?
+    # changed in 2.9.14, see https://github.com/sparklemotion/nokogiri/releases/tag/v1.13.5
+    # then reverted in 2.10.0, see https://gitlab.gnome.org/GNOME/libxml2/-/issues/380
+    Nokogiri.method(:uses_libxml?).arity == -1 && Nokogiri.uses_libxml?("= 2.9.14")
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: rails-html-sanitizer
 version: !ruby/object:Gem::Version
-  version: 1.4.3
+  version: 1.4.4
 platform: ruby
 authors:
 - Rafael Mendonça França
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2022-06-09 00:00:00.000000000 Z
+date: 2022-12-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: loofah
@@ -17,14 +17,20 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.3'
+        version: '2.19'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 2.19.1
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '2.3'
+        version: '2.19'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 2.19.1
 - !ruby/object:Gem::Dependency
   name: bundler
   requirement: !ruby/object:Gem::Requirement
@@ -103,9 +109,9 @@ licenses:
 - MIT
 metadata:
   bug_tracker_uri: https://github.com/rails/rails-html-sanitizer/issues
-  changelog_uri: https://github.com/rails/rails-html-sanitizer/blob/v1.4.3/CHANGELOG.md
-  documentation_uri: https://www.rubydoc.info/gems/rails-html-sanitizer/1.4.3
-  source_code_uri: https://github.com/rails/rails-html-sanitizer/tree/v1.4.3
+  changelog_uri: https://github.com/rails/rails-html-sanitizer/blob/v1.4.4/CHANGELOG.md
+  documentation_uri: https://www.rubydoc.info/gems/rails-html-sanitizer/1.4.4
+  source_code_uri: https://github.com/rails/rails-html-sanitizer/tree/v1.4.4
 post_install_message: 
 rdoc_options: []
 require_paths:
@@ -121,7 +127,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.3.5
+rubygems_version: 3.3.7
 signing_key: 
 specification_version: 4
 summary: This gem is responsible to sanitize HTML fragments in Rails applications.