nokogiri 1.7.2-java → 1.8.0-java

data/lib/nokogiri/xml/node_set.rb

@@ -11,6 +11,8 @@ module Nokogiri
       # The Document this NodeSet is associated with
       attr_accessor :document
 
+      alias :clone :dup
+
       # Create a NodeSet with +document+ defaulting to +list+
       def initialize document, list = []
         @document = document
@@ -71,9 +73,10 @@ module Nokogiri
       # For more information see Nokogiri::XML::Searchable#css
       def css *args
         rules, handler, ns, _ = extract_params(args)
+        paths = css_rules_to_xpath(rules, ns)
 
         inject(NodeSet.new(document)) do |set, node|
-          set += css_internal node, rules, handler, ns
+          set + xpath_internal(node, paths, handler, ns, nil)
         end
       end
 
@@ -88,7 +91,7 @@ module Nokogiri
         paths, handler, ns, binds = extract_params(args)
 
         inject(NodeSet.new(document)) do |set, node|
-          set += node.xpath(*(paths + [ns, handler, binds].compact))
+          set + xpath_internal(node, paths, handler, ns, binds)
         end
       end
 
@@ -290,7 +293,11 @@ module Nokogiri
       # Returns a new NodeSet containing all the children of all the nodes in
       # the NodeSet
       def children
-        inject(NodeSet.new(document)) { |set, node| set += node.children }
+        node_set = NodeSet.new(document)
+        each do |node|
+          node.children.each { |n| node_set.push(n) }
+        end
+        node_set
       end
 
       ###
@@ -312,11 +319,9 @@ module Nokogiri
 
       alias :+ :|
 
-      private
+      # @private
+      IMPLIED_XPATH_CONTEXTS = [ './/'.freeze, 'self::'.freeze ].freeze # :nodoc:
 
-      def implied_xpath_contexts # :nodoc:
-        [".//", "self::"]
-      end
     end
   end
 end

data/lib/nokogiri/xml/sax/parser.rb

@@ -68,8 +68,7 @@ module Nokogiri
 
         # Create a new Parser with +doc+ and +encoding+
         def initialize doc = Nokogiri::XML::SAX::Document.new, encoding = 'UTF-8'
-          check_encoding(encoding)
-          @encoding = encoding
+          @encoding = check_encoding(encoding)
           @document = doc
           @warned   = false
         end
@@ -88,9 +87,8 @@ module Nokogiri
         ###
         # Parse given +io+
         def parse_io io, encoding = 'ASCII'
-          check_encoding(encoding)
-          @encoding = encoding
-          ctx = ParserContext.io(io, ENCODINGS[encoding])
+          @encoding = check_encoding(encoding)
+          ctx = ParserContext.io(io, ENCODINGS[@encoding])
           yield ctx if block_given?
           ctx.parse_with self
         end
@@ -114,8 +112,9 @@ module Nokogiri
 
         private
         def check_encoding(encoding)
-          encoding.upcase!
-          raise ArgumentError.new("'#{encoding}' is not a valid encoding") unless ENCODINGS[encoding]
+          encoding.upcase.tap do |enc|
+            raise ArgumentError.new("'#{enc}' is not a valid encoding") unless ENCODINGS[enc]
+          end
         end
       end
     end

data/lib/nokogiri/xml/searchable.rb

@@ -149,30 +149,9 @@ module Nokogiri
       #   }.new)
       #
       def xpath *args
-        return NodeSet.new(document) unless document
-
         paths, handler, ns, binds = extract_params(args)
 
-        sets = paths.map do |path|
-          ctx = XPathContext.new(self)
-          ctx.register_namespaces(ns)
-          path = path.gsub(/xmlns:/, ' :') unless Nokogiri.uses_libxml?
-
-          binds.each do |key,value|
-            ctx.register_variable key.to_s, value
-          end if binds
-
-          ctx.evaluate(path, handler)
-        end
-        return sets.first if sets.length == 1
-
-        NodeSet.new(document) do |combined|
-          sets.each do |set|
-            set.each do |node|
-              combined << node
-            end
-          end
-        end
+        xpath_internal self, paths, handler, ns, binds
       end
 
       ##
@@ -189,12 +168,42 @@ module Nokogiri
       private
 
       def css_internal node, rules, handler, ns
-        xpaths = rules.map { |rule| xpath_query_from_css_rule(rule, ns) }
-        node.xpath(*(xpaths + [ns, handler].compact))
+        xpath_internal node, css_rules_to_xpath(rules, ns), handler, ns, nil
+      end
+
+      def xpath_internal node, paths, handler, ns, binds
+        document = node.document
+        return NodeSet.new(document) unless document
+
+        if paths.length == 1
+          return xpath_impl(node, paths.first, handler, ns, binds)
+        end
+
+        NodeSet.new(document) do |combined|
+          paths.each do |path|
+            xpath_impl(node, path, handler, ns, binds).each { |set| combined << set }
+          end
+        end
+      end
+
+      def xpath_impl node, path, handler, ns, binds
+        ctx = XPathContext.new(node)
+        ctx.register_namespaces(ns)
+        path = path.gsub(/xmlns:/, ' :') unless Nokogiri.uses_libxml?
+
+        binds.each do |key,value|
+          ctx.register_variable key.to_s, value
+        end if binds
+
+        ctx.evaluate(path, handler)
+      end
+
+      def css_rules_to_xpath(rules, ns)
+        rules.map { |rule| xpath_query_from_css_rule(rule, ns) }
       end
 
       def xpath_query_from_css_rule rule, ns
-        implied_xpath_contexts.map do |implied_xpath_context|
+        self.class::IMPLIED_XPATH_CONTEXTS.map do |implied_xpath_context|
           CSS.xpath_for(rule.to_s, :prefix => implied_xpath_context, :ns => ns)
         end.join(' | ')
       end

data/lib/nokogiri/xml/syntax_error.rb

@@ -40,7 +40,30 @@ module Nokogiri
       end
 
       def to_s
-        super.chomp
+        message = super.chomp
+        [location_to_s, level_to_s, message].
+          compact.join(": ").
+          force_encoding(message.encoding)
+      end
+
+      private
+
+      def level_to_s
+        case level
+        when 3 then "FATAL"
+        when 2 then "ERROR"
+        when 1 then "WARNING"
+        else nil
+        end
+      end
+
+      def nil_or_zero?(attribute)
+        attribute.nil? || attribute.zero?
+      end
+
+      def location_to_s
+        return nil if nil_or_zero?(line) && nil_or_zero?(column)
+        "#{line}:#{column}"
       end
     end
   end

data/test/decorators/test_slop.rb

@@ -9,10 +9,13 @@ module Nokogiri
           <description>this is the foo thing</description>
         </item>
       eoxml
+
       assert doc.item.respond_to?(:title)
       assert_equal 'foo', doc.item.title.text
+
       assert doc.item.respond_to?(:_description), 'should have description'
-      assert 'this is the foo thing', doc.item._description.text
+      assert_equal 'this is the foo thing', doc.item._description.text
+
       assert !doc.item.respond_to?(:foo)
       assert_raise(NoMethodError) { doc.item.foo }
     end

data/test/helper.rb

@@ -7,7 +7,17 @@ require 'tempfile'
 require 'pp'
 
 require 'nokogiri'
+
 if ENV['TEST_NOKOGIRI_WITH_LIBXML_RUBY']
+  #
+  #  if you'd like to test with the libxml-ruby gem loaded, it's
+  #  recommended that you set
+  #
+  #    BUNDLE_GEMFILE=Gemfile-libxml-ruby
+  #
+  #  which will a) bundle that gem, and b) set the appropriate env var to
+  #  trigger this block
+  #
   require 'libxml'
   warn "#{__FILE__}:#{__LINE__}: loaded libxml-ruby '#{LibXML::XML::VERSION}'"
 end

data/test/html/sax/test_parser.rb

@@ -132,6 +132,33 @@ module Nokogiri
           ], @parser.document.start_elements
         end
 
+        HTML_WITH_BR_TAG = <<-EOF
+        <html>
+          <head></head>
+          <body>
+            <div>
+              hello
+              <br>
+            </div>
+
+            <div>
+              hello again
+            </div>
+          </body>
+        </html>
+        EOF
+
+        def test_parsing_dom_error_from_string
+          @parser.parse(HTML_WITH_BR_TAG)
+          assert_equal 6, @parser.document.start_elements.length
+        end
+
+        def test_parsing_dom_error_from_io
+          @parser.parse(StringIO.new(HTML_WITH_BR_TAG))
+          assert_equal 6, @parser.document.start_elements.length
+        end
+
+
         def test_empty_processing_instruction
           @parser.parse_memory("<strong>this will segfault<?strong>")
         end

data/test/html/test_document.rb

@@ -15,7 +15,7 @@ module Nokogiri
       end
 
       def test_does_not_fail_with_illformatted_html
-        doc = Nokogiri::HTML('"</html>";'.force_encoding(Encoding::BINARY))
+        doc = Nokogiri::HTML('"</html>";'.dup.force_encoding(Encoding::BINARY))
         assert_not_nil doc
       end
 
@@ -696,6 +696,17 @@ eohtml
         node = html.xpath("//style").first
         assert_equal("tr > div { display:block; }", node.inner_html)
       end
+
+      it "does not fail when converting to_html using explicit encoding" do
+        html_fragment=<<-eos
+  <img width="16" height="16" src="images/icon.gif" border="0" alt="Inactive hide details for &quot;User&quot; ---19/05/2015 12:55:29---Provvediamo subito nell&#8217;integrare">
+        eos
+        doc = Nokogiri::HTML(html_fragment, nil, 'ISO-8859-1')
+        html = doc.to_html
+        assert html.index("src=\"images/icon.gif\"")
+        assert_equal 'ISO-8859-1', html.encoding.name
+      end
+
     end
   end
 end

data/test/html/test_document_encoding.rb

@@ -128,9 +128,7 @@ module Nokogiri
           assert_equal(title, doc_from_string_enc.at('//title/text()').text)
           assert_equal(title, doc_from_string.at('//title/text()').text)
           assert_equal(title, doc_from_file_enc.at('//title/text()').text)
-          unless Nokogiri.jruby? && file == ENCODING_HTML_FILE
-            assert_equal(title, doc_from_file.at('//title/text()').text)
-          end
+          assert_equal(title, doc_from_file.at('//title/text()').text)
 
           evil = (0..72).map { |i| '超' * i + '悪い事を構想中。' }
 

data/test/html/test_document_fragment.rb

@@ -234,6 +234,9 @@ module Nokogiri
       end
 
       def test_element_children_counts
+        if Nokogiri.uses_libxml? && Nokogiri::VERSION_INFO['libxml']['loaded'] <= "2.9.1"
+          skip "#elements doesn't work in 2.9.1, see 1793a5a for history"
+        end
         doc = Nokogiri::HTML::DocumentFragment.parse("   <div>  </div>\n   ")
         assert_equal 1, doc.element_children.count
       end

data/test/xml/sax/test_push_parser.rb

@@ -151,6 +151,54 @@ module Nokogiri
           assert_equal "Gau\337", @parser.document.data.join
           assert_equal [["r"]], @parser.document.end_elements
         end
+
+        def test_replace_entities_attribute_behavior
+          if Nokogiri.uses_libxml?
+            # initially false
+            assert_equal false, @parser.replace_entities
+
+            # can be set to true
+            @parser.replace_entities = true
+            assert_equal true, @parser.replace_entities
+
+            # can be set to false
+            @parser.replace_entities = false
+            assert_equal false, @parser.replace_entities
+          else
+            # initially true
+            assert_equal true, @parser.replace_entities
+
+            # ignore attempts to set to false
+            @parser.replace_entities = false # TODO: should we raise an exception here?
+            assert_equal true, @parser.replace_entities
+          end
+        end
+
+        def test_untouched_entities
+          skip("entities are always replaced in pure Java version") if Nokogiri.jruby?
+          @parser.<<(<<-eoxml)
+            <p id="asdf&amp;asdf">
+              <!-- This is a comment -->
+              Paragraph 1 &amp; 2
+            </p>
+          eoxml
+          @parser.finish
+          assert_equal [["p", [["id", "asdf&#38;asdf"]]]], @parser.document.start_elements
+          assert_equal "Paragraph 1 & 2", @parser.document.data.join.strip
+        end
+
+        def test_replaced_entities
+          @parser.replace_entities = true
+          @parser.<<(<<-eoxml)
+            <p id="asdf&amp;asdf">
+              <!-- This is a comment -->
+              Paragraph 1 &amp; 2
+            </p>
+          eoxml
+          @parser.finish
+          assert_equal [["p", [["id", "asdf&asdf"]]]], @parser.document.start_elements
+          assert_equal "Paragraph 1 & 2", @parser.document.data.join.strip
+        end
       end
     end
   end

data/test/xml/test_attr.rb

@@ -11,6 +11,13 @@ module Nokogiri
         }
       end
 
+      def test_new_raises_argerror_on_nondocument
+        document = Nokogiri::XML "<root><foo/></root>"
+        assert_raises ArgumentError do
+          Nokogiri::XML::Attr.new document.at_css("foo"), "bar"
+        end
+      end
+
       def test_content=
         xml = Nokogiri::XML.parse(File.read(XML_FILE), XML_FILE)
         address = xml.xpath('//address')[3]

data/test/xml/test_document.rb

@@ -221,7 +221,7 @@ module Nokogiri
       end
 
       def test_pp
-        out = StringIO.new('')
+        out = StringIO.new(String.new)
         ::PP.pp @xml, out
         assert_operator out.string.length, :>, 0
       end

data/test/xml/test_document_fragment.rb

@@ -136,6 +136,23 @@ module Nokogiri
         assert_equal expected, search_xpath
       end
 
+      def test_fragment_css_search_with_whitespace_and_node_removal
+        # The same xml without leading whitespace in front of the first line
+        # does not expose the error. Putting both nodes on the same line
+        # instead also fixes the crash.
+        fragment = Nokogiri::XML::DocumentFragment.parse <<-EOXML
+          <p id="content">hi</p> x <!--y--> <p>another paragraph</p>
+        EOXML
+        children = fragment.css('p')
+        assert_equal 2, children.length
+        # removing the last node instead does not yield the error. Probably the
+        # node removal leaves around two consecutive text nodes which make the
+        # css search crash?
+        children.first.remove
+        assert_equal 1, fragment.xpath('.//p | self::p').length
+        assert_equal 1, fragment.css('p').length
+      end
+
       def test_fragment_search_three_ways
         frag = Nokogiri::XML::Document.new.fragment '<p id="content">foo</p><p id="content">bar</p>'
         expected = frag.xpath('./*[@id = "content"]')
@@ -219,6 +236,16 @@ module Nokogiri
         assert fragment.children.respond_to?(:awesome!), fragment.children.class
       end
 
+      def test_decorator_is_applied_to_empty_set
+        x = Module.new do
+          def awesome!
+          end
+        end
+        util_decorate(@xml, x)
+        fragment = Nokogiri::XML::DocumentFragment.new(@xml, "")
+        assert fragment.children.respond_to?(:awesome!), fragment.children.class
+      end
+
       def test_add_node_to_doc_fragment_segfault
         frag = Nokogiri::XML::DocumentFragment.new(@xml, '<p>hello world</p>')
         Nokogiri::XML::Comment.new(frag,'moo')

data/test/xml/test_entity_reference.rb

@@ -141,7 +141,7 @@ EOF
         end
         assert_kind_of Nokogiri::XML::EntityReference, doc.xpath('//body').first.children.first
         if Nokogiri.uses_libxml?
-          assert_equal ["Entity 'bar' not defined"], doc.errors.map(&:to_s)
+          assert_equal ["5:14: ERROR: Entity 'bar' not defined"], doc.errors.map(&:to_s)
         end
       end
 
@@ -159,7 +159,7 @@ EOF
         end
         assert_kind_of Nokogiri::XML::EntityReference, doc.xpath('//body').first.children.first
         if Nokogiri.uses_libxml?
-          assert_equal ["Attempt to load network entity http://foo.bar.com/", "Entity 'bar' not defined"], doc.errors.map(&:to_s)
+          assert_equal ["ERROR: Attempt to load network entity http://foo.bar.com/", "4:34: ERROR: Entity 'bar' not defined"], doc.errors.map(&:to_s)
         else
           assert_equal ["Attempt to load network entity http://foo.bar.com/"], doc.errors.map(&:to_s)
         end