Merge pull request #49 from taleksei/29908

feat: add options for sanitize outer links in css

lib/string_tools.rb

@@ -155,13 +155,21 @@ class Base
 
       TAGS_WITHOUT_ATTRIBUTES = %w(b strong i em sup sub ul ol li blockquote br tr u caption thead s).freeze
 
+      # Public: Sanitize string
+      # str - String for sanitize
+      # attrs - Hash, custom attributes, defaults empty hash
+      #   remove_contents - Set of string, tags to be removed
+      #   protocols - Array of string, protocols using in css properties urls
       def sanitize(str, attrs = {})
         # для корректного обрезания utf строчек режем через mb_chars
         # для защиты от перегрузки парсера пропускаем максимум 1 мегабайт текста
         # длина русского символа в utf-8 - 2 байта, 1Мб/2б = 524288 = 2**19 символов
         # длина по символам с перестраховкой, т.к. латинские символы(теги, например) занимают 1 байт
         str = str.mb_chars.slice(0..(2**19)).to_s
 
+        remove_contents = attrs.delete(:remove_contents)
+        protocols = attrs.delete(:protocols) || []
+
         # Мерджим добавочные теги и атрибуты
         attributes = TAGS_WITH_ATTRIBUTES.merge(attrs)
         elements = attributes.keys | TAGS_WITHOUT_ATTRIBUTES
@@ -173,8 +181,8 @@ def sanitize(str, attrs = {})
           str,
           :attributes => attributes,
           :elements => elements,
-          :css => {:properties => Sanitize::Config::RELAXED[:css][:properties]},
-          :remove_contents => %w(style script),
+          :css => {:properties => Sanitize::Config::RELAXED[:css][:properties], protocols: protocols},
+          :remove_contents => remove_contents || Set['style', 'script'],
           :allow_comments => false,
           :transformers => transformers
         )
@@ -191,6 +199,7 @@ def call(env)
           normalize_link node, 'href'
         when 'img'
           normalize_link node, 'src'
+          remove_links node, 'alt'
         end
       end
 
@@ -202,6 +211,14 @@ def normalize_link(node, attr_name)
       rescue Addressable::URI::InvalidURIError
         node.swap node.children
       end
+
+      def remove_links(node, attr_name)
+        return unless node[attr_name]
+
+        node[attr_name] = node[attr_name].gsub(URI::DEFAULT_PARSER.make_regexp, '').squish
+
+        node.remove_attribute(attr_name) if node[attr_name].empty?
+      end
     end
 
     class IframeNormalizer

spec/string_tools_spec.rb

@@ -56,6 +56,42 @@
         to eq('<iframe width="123" height="456" src="https://www.youtube.com/embed/abc" frameborder="0"></iframe>')
     end
 
+    it 'removes outer link from css when protocols given' do
+      origin_str = '<div style="background-image: url(http://i54.tinypic.com/4zuxif.jpg)"></div>'
+      sanitized_string = described_class.sanitize(origin_str)
+      expect(sanitized_string).to eq('<div></div>')
+    end
+
+    it 'do not removes outer link from css when protocols given' do
+      origin_str = '<div style="background-image: url(http://i54.tinypic.com/4zuxif.jpg)"></div>'
+      sanitized_string = described_class.sanitize(origin_str, protocols: %w[http https])
+      expect(sanitized_string).to eq('<div style="background-image: url(http://i54.tinypic.com/4zuxif.jpg)"></div>')
+    end
+
+    it 'removes style content' do
+      origin_str = '<style type="text/css">body{color: red;}</style>'
+      sanitized_string = described_class.sanitize(origin_str)
+      expect(sanitized_string).to eq('')
+    end
+
+    it 'do not removes style content' do
+      origin_str = '<style type="text/css">body{color: red;}</style>'
+      sanitized_string = described_class.sanitize(origin_str, 'style' => %w(type), remove_contents: Set['script'])
+      expect(sanitized_string).to eq('<style type="text/css">body{color: red;}</style>')
+    end
+
+    it 'removes links in alt attribute of img tag' do
+      origin_str = '<img scr="http://test.test" alt="http://test.test test https://test.test alt">'
+      sanitized_string = described_class.sanitize(origin_str, 'img' => %w(scr alt))
+      expect(sanitized_string).to eq('<img scr="http://test.test" alt="test alt">')
+    end
+
+    it 'removes alt attribute of img tag if empty value' do
+      origin_str = '<img scr="http://test.test" alt="http://test.test">'
+      sanitized_string = described_class.sanitize(origin_str, 'img' => %w(scr alt))
+      expect(sanitized_string).to eq('<img scr="http://test.test">')
+    end
+
     context 'multiple invocations of the method' do
       it 'does not mess up default config' do
         origin_str = '<p style="text-align: center;" title="foobar"></p>'