nokogiri 1.7.2 → 1.8.0

data/ext/nokogiri/xml_sax_push_parser.c

@@ -3,7 +3,7 @@
 static void deallocate(xmlParserCtxtPtr ctx)
 {
   NOKOGIRI_DEBUG_START(ctx);
-  if(ctx != NULL) {
+  if (ctx != NULL) {
     NOKOGIRI_SAX_TUPLE_DESTROY(ctx->userData);
     xmlFreeParserCtxt(ctx);
   }
@@ -30,12 +30,12 @@ static VALUE native_write(VALUE self, VALUE _chunk, VALUE _last_chunk)
 
   Data_Get_Struct(self, xmlParserCtxt, ctx);
 
-  if(Qnil != _chunk) {
+  if (Qnil != _chunk) {
     chunk = StringValuePtr(_chunk);
     size = (int)RSTRING_LEN(_chunk);
   }
 
-  if(xmlParseChunk(ctx, chunk, size, Qtrue == _last_chunk ? 1 : 0)) {
+  if (xmlParseChunk(ctx, chunk, size, Qtrue == _last_chunk ? 1 : 0)) {
     if (!(ctx->options & XML_PARSE_RECOVER)) {
       xmlErrorPtr e = xmlCtxtGetLastError(ctx);
       Nokogiri_error_raise(NULL, e);
@@ -59,17 +59,18 @@ static VALUE initialize_native(VALUE self, VALUE _xml_sax, VALUE _filename)
 
   Data_Get_Struct(_xml_sax, xmlSAXHandler, sax);
 
-  if(_filename != Qnil) filename = StringValueCStr(_filename);
+  if (_filename != Qnil) { filename = StringValueCStr(_filename); }
 
   ctx = xmlCreatePushParserCtxt(
-      sax,
-      NULL,
-      NULL,
-      0,
-      filename
-  );
-  if(ctx == NULL)
+          sax,
+          NULL,
+          NULL,
+          0,
+          filename
+        );
+  if (ctx == NULL) {
     rb_raise(rb_eRuntimeError, "Could not create a parser context");
+  }
 
   ctx->userData = NOKOGIRI_SAX_TUPLE_NEW(ctx, self);
 
@@ -91,12 +92,53 @@ static VALUE set_options(VALUE self, VALUE options)
   xmlParserCtxtPtr ctx;
   Data_Get_Struct(self, xmlParserCtxt, ctx);
 
-  if (xmlCtxtUseOptions(ctx, (int)NUM2INT(options)) != 0)
+  if (xmlCtxtUseOptions(ctx, (int)NUM2INT(options)) != 0) {
     rb_raise(rb_eRuntimeError, "Cannot set XML parser context options");
+  }
 
   return Qnil;
 }
 
+/*
+ * call-seq:
+ *  replace_entities
+ *
+ * Should this parser replace entities?  & will get converted to '&' if
+ * set to true
+ */
+static VALUE get_replace_entities(VALUE self)
+{
+  xmlParserCtxtPtr ctx;
+  Data_Get_Struct(self, xmlParserCtxt, ctx);
+
+  if (0 == ctx->replaceEntities) {
+    return Qfalse;
+  } else {
+    return Qtrue;
+  }
+}
+
+/*
+ * call-seq:
+ *  replace_entities=(boolean)
+ *
+ * Should this parser replace entities?  & will get converted to '&' if
+ * set to true
+ */
+static VALUE set_replace_entities(VALUE self, VALUE value)
+{
+  xmlParserCtxtPtr ctx;
+  Data_Get_Struct(self, xmlParserCtxt, ctx);
+
+  if (Qfalse == value) {
+    ctx->replaceEntities = 0;
+  } else {
+    ctx->replaceEntities = 1;
+  }
+
+  return value;
+}
+
 VALUE cNokogiriXmlSaxPushParser ;
 void init_xml_sax_push_parser()
 {
@@ -112,4 +154,6 @@ void init_xml_sax_push_parser()
   rb_define_private_method(klass, "native_write", native_write, 2);
   rb_define_method(klass, "options", get_options, 0);
   rb_define_method(klass, "options=", set_options, 1);
+  rb_define_method(klass, "replace_entities", get_replace_entities, 0);
+  rb_define_method(klass, "replace_entities=", set_replace_entities, 1);
 }

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

@@ -36,6 +36,16 @@ module Nokogiri
           ctx.parse_with self
         end
 
+        ###
+        # Parse given +io+
+        def parse_io io, encoding = 'UTF-8'
+          check_encoding(encoding)
+          @encoding = encoding
+          ctx = ParserContext.io(io, ENCODINGS[encoding])
+          yield ctx if block_given?
+          ctx.parse_with self
+        end
+
         ###
         # Parse a file with +filename+
         def parse_file filename, encoding = 'UTF-8'

data/lib/nokogiri/version.rb

@@ -1,6 +1,6 @@
 module Nokogiri
   # The version of Nokogiri you are using
-  VERSION = '1.7.2'
+  VERSION = '1.8.0'
 
   class VersionInfo # :nodoc:
     def jruby?
@@ -12,9 +12,10 @@ module Nokogiri
     end
 
     def loaded_parser_version
-      LIBXML_PARSER_VERSION.scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first.collect{ |j|
-        j.to_i
-      }.join(".")
+      LIBXML_PARSER_VERSION.
+        scan(/^(\d+)(\d\d)(\d\d)(?!\d)/).first.
+        collect(&:to_i).
+        join(".")
     end
 
     def compiled_parser_version

data/lib/nokogiri/xml/document.rb

@@ -33,10 +33,11 @@ module Nokogiri
       # +block+ (optional) is passed a configuration object on which
       # parse options may be set.
       #
-      # When parsing untrusted documents, it's recommended that the
-      # +nonet+ option be used, as shown in this example code:
-      #
-      #   Nokogiri::XML::Document.parse(xml_string) { |config| config.nonet }
+      # By default, Nokogiri treats documents as untrusted, and so
+      # does not attempt to load DTDs or access the network. See
+      # Nokogiri::XML::ParseOptions for a complete list of options;
+      # and that module's DEFAULT_XML constant for what's set (and not
+      # set) by default.
       #
       # Nokogiri.XML() is a convenience method which will call this method.
       #
@@ -239,10 +240,10 @@ module Nokogiri
       undef_method :namespace_definitions, :line, :add_namespace
 
       def add_child node_or_tags
-        raise "Document already has a root node" if root && root.name != 'nokogiri_text_wrapper'
+        raise "A document may not have multiple root nodes." if (root && root.name != 'nokogiri_text_wrapper') && !(node_or_tags.comment? || node_or_tags.processing_instruction?)
         node_or_tags = coerce(node_or_tags)
         if node_or_tags.is_a?(XML::NodeSet)
-          raise "Document cannot have multiple root nodes" if node_or_tags.size > 1
+          raise "A document may not have multiple root nodes." if node_or_tags.size > 1
           super(node_or_tags.first)
         else
           super
@@ -273,9 +274,8 @@ module Nokogiri
           (string_or_io.respond_to?(:eof?) && string_or_io.eof?)
       end
 
-      def implied_xpath_contexts # :nodoc:
-        ["//"]
-      end
+      # @private
+      IMPLIED_XPATH_CONTEXTS = [ '//'.freeze ].freeze # :nodoc:
 
       def inspect_attributes
         [:name, :children]

data/lib/nokogiri/xml/node.rb

@@ -154,7 +154,7 @@ module Nokogiri
       def prepend_child node_or_tags
         if first = children.first
           # Mimic the error add_child would raise.
-          raise RuntimeError, "Document already has a root node" if document? && !node_or_tags.processing_instruction?
+          raise RuntimeError, "Document already has a root node" if document? && !(node_or_tags.comment? || node_or_tags.processing_instruction?)
           first.__send__(:add_sibling, :previous, node_or_tags)
         else
           add_child(node_or_tags)
@@ -172,6 +172,7 @@ module Nokogiri
         add_child node_or_tags
         self
       end
+
       ###
       # Insert +node_or_tags+ before this Node (as a sibling).
       # +node_or_tags+ can be a Nokogiri::XML::Node, a ::DocumentFragment, a ::NodeSet, or a string containing markup.
@@ -180,7 +181,7 @@ module Nokogiri
       #
       # Also see related method +before+.
       def add_previous_sibling node_or_tags
-        raise ArgumentError.new("A document may not have multiple root nodes.") if (parent && parent.document?) && !node_or_tags.processing_instruction?
+        raise ArgumentError.new("A document may not have multiple root nodes.") if (parent && parent.document?) && !(node_or_tags.comment? || node_or_tags.processing_instruction?)
 
         add_sibling :previous, node_or_tags
       end
@@ -193,7 +194,7 @@ module Nokogiri
       #
       # Also see related method +after+.
       def add_next_sibling node_or_tags
-        raise ArgumentError.new("A document may not have multiple root nodes.") if (parent && parent.document?) && !node_or_tags.processing_instruction?
+        raise ArgumentError.new("A document may not have multiple root nodes.") if (parent && parent.document?) && !(node_or_tags.comment? || node_or_tags.processing_instruction?)
 
         add_sibling :next, node_or_tags
       end
@@ -622,7 +623,7 @@ module Nokogiri
         encoding = options[:encoding] || document.encoding
         options[:encoding] = encoding
 
-        outstring = ""
+        outstring = String.new
         if encoding && outstring.respond_to?(:force_encoding)
           outstring.force_encoding(Encoding.find(encoding))
         end
@@ -818,9 +819,8 @@ Requires a Node, NodeSet or String argument, and cannot accept a #{data.class}.
         EOERR
       end
 
-      def implied_xpath_contexts # :nodoc:
-        [".//"]
-      end
+      # @private
+      IMPLIED_XPATH_CONTEXTS = [ './/'.freeze ].freeze # :nodoc:
 
       def add_child_node_and_reparent_attrs node # :nodoc:
         add_child_node node

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