nokogiri 1.6.5 → 1.6.6.1

data/ext/nokogiri/xml_comment.c

@@ -1,10 +1,13 @@
 #include <xml_comment.h>
 
+static ID document_id ;
+
 /*
  * call-seq:
- *  new(document, content)
+ *  new(document_or_node, content)
  *
- * Create a new Comment element on the +document+ with +content+
+ * Create a new Comment element on the +document+ with +content+.
+ * Alternatively, if a +node+ is passed, the +node+'s document is used.
  */
 static VALUE new(int argc, VALUE *argv, VALUE klass)
 {
@@ -17,6 +20,16 @@ static VALUE new(int argc, VALUE *argv, VALUE klass)
 
   rb_scan_args(argc, argv, "2*", &document, &content, &rest);
 
+  if (rb_obj_is_kind_of(document, cNokogiriXmlNode))
+  {
+    document = rb_funcall(document, document_id, 0);
+  }
+  else if (   !rb_obj_is_kind_of(document, cNokogiriXmlDocument)
+           && !rb_obj_is_kind_of(document, cNokogiriXmlDocumentFragment))
+  {
+    rb_raise(rb_eArgError, "first argument must be a XML::Document or XML::Node");
+  }
+
   Data_Get_Struct(document, xmlDoc, xml_doc);
 
   node = xmlNewDocComment(
@@ -51,4 +64,6 @@ void init_xml_comment()
   cNokogiriXmlComment = klass;
 
   rb_define_singleton_method(klass, "new", new, -1);
+
+  document_id = rb_intern("document");
 }

data/ext/nokogiri/xml_node.c

@@ -14,7 +14,7 @@ static void debug_node_dealloc(xmlNodePtr x)
 
 static void mark(xmlNodePtr node)
 {
-  xmlNodePtr doc = node->doc;
+  xmlDocPtr doc = node->doc;
   if(doc->type == XML_DOCUMENT_NODE || doc->type == XML_HTML_DOCUMENT_NODE) {
     if(DOC_RUBY_OBJECT_TEST(doc)) {
       rb_gc_mark(DOC_RUBY_OBJECT(doc));
@@ -234,7 +234,15 @@ ok:
      *  reparent the actual reparentee, so we reparent a duplicate.
      */
     nokogiri_root_node(reparentee);
-    if (!(reparentee = xmlDocCopyNode(reparentee, pivot->doc, 1))) {
+
+    xmlResetLastError();
+    xmlSetStructuredErrorFunc((void *)rb_iv_get(DOC_RUBY_OBJECT(pivot->doc), "@errors"), Nokogiri_error_array_pusher);
+
+    reparentee = xmlDocCopyNode(reparentee, pivot->doc, 1) ;
+
+    xmlSetStructuredErrorFunc(NULL, NULL);
+
+    if (! reparentee) {
       rb_raise(rb_eRuntimeError, "Could not reparent node (xmlDocCopyNode)");
     }
   }
@@ -473,7 +481,13 @@ static VALUE duplicate_node(int argc, VALUE *argv, VALUE self)
 
   Data_Get_Struct(self, xmlNode, node);
 
+  xmlResetLastError();
+  xmlSetStructuredErrorFunc(NULL, Nokogiri_error_silencer);
+
   dup = xmlDocCopyNode(node, node->doc, (int)NUM2INT(level));
+
+  xmlSetStructuredErrorFunc(NULL, NULL);
+
   if(dup == NULL) return Qnil;
 
   nokogiri_root_node(dup);
@@ -1012,7 +1026,7 @@ static VALUE node_type(VALUE self)
  *
  * Set the content for this Node
  */
-static VALUE native_content(VALUE self, VALUE content)
+static VALUE set_native_content(VALUE self, VALUE content)
 {
   xmlNodePtr node, child, next ;
   Data_Get_Struct(self, xmlNode, node);
@@ -1035,7 +1049,7 @@ static VALUE native_content(VALUE self, VALUE content)
  *
  * Returns the content for this Node
  */
-static VALUE get_content(VALUE self)
+static VALUE get_native_content(VALUE self)
 {
   xmlNodePtr node;
   xmlChar * content;
@@ -1051,6 +1065,50 @@ static VALUE get_content(VALUE self)
   return Qnil;
 }
 
+/*
+ * call-seq:
+ *  lang=
+ *
+ * Set the language of a node, i.e. the values of the xml:lang attribute.
+ */
+static VALUE set_lang(VALUE self_rb, VALUE lang_rb)
+{
+  xmlNodePtr self ;
+  xmlChar* lang ;
+
+  Data_Get_Struct(self_rb, xmlNode, self);
+  lang = (xmlChar*)StringValuePtr(lang_rb);
+
+  xmlNodeSetLang(self, lang);
+
+  return Qnil ;
+}
+
+/*
+ * call-seq:
+ *  lang
+ *
+ * Searches the language of a node, i.e. the values of the xml:lang attribute or
+ * the one carried by the nearest ancestor.
+ */
+static VALUE get_lang(VALUE self_rb)
+{
+  xmlNodePtr self ;
+  xmlChar* lang ;
+  VALUE lang_rb ;
+
+  Data_Get_Struct(self_rb, xmlNode, self);
+
+  lang = xmlNodeGetLang(self);
+  if (lang) {
+    lang_rb = NOKOGIRI_STR_NEW2(lang);
+    xmlFree(lang);
+    return lang_rb ;
+  }
+
+  return Qnil ;
+}
+
 /* :nodoc: */
 static VALUE add_child(VALUE self, VALUE new_child)
 {
@@ -1561,7 +1619,6 @@ void init_xml_node()
   rb_define_method(klass, "next_element", next_element, 0);
   rb_define_method(klass, "previous_element", previous_element, 0);
   rb_define_method(klass, "node_type", node_type, 0);
-  rb_define_method(klass, "content", get_content, 0);
   rb_define_method(klass, "path", path, 0);
   rb_define_method(klass, "key?", key_eh, 1);
   rb_define_method(klass, "namespaced_key?", namespaced_key_eh, 2);
@@ -1581,7 +1638,10 @@ void init_xml_node()
   rb_define_method(klass, "create_external_subset", create_external_subset, 3);
   rb_define_method(klass, "pointer_id", pointer_id, 0);
   rb_define_method(klass, "line", line, 0);
-  rb_define_method(klass, "native_content=", native_content, 1);
+  rb_define_method(klass, "content", get_native_content, 0);
+  rb_define_method(klass, "native_content=", set_native_content, 1);
+  rb_define_method(klass, "lang", get_lang, 0);
+  rb_define_method(klass, "lang=", set_lang, 1);
 
   rb_define_private_method(klass, "process_xincludes", process_xincludes, 1);
   rb_define_private_method(klass, "in_context", in_context, 2);

data/ext/nokogiri/xml_syntax_error.c

@@ -7,6 +7,10 @@ void Nokogiri_error_array_pusher(void * ctx, xmlErrorPtr error)
   rb_ary_push(list,  Nokogiri_wrap_xml_syntax_error(error));
 }
 
+void Nokogiri_error_silencer(void * ctx, xmlErrorPtr error)
+{
+}
+
 void Nokogiri_error_raise(void * ctx, xmlErrorPtr error)
 {
   rb_exc_raise(Nokogiri_wrap_xml_syntax_error(error));

data/ext/nokogiri/xml_syntax_error.h

@@ -6,6 +6,7 @@
 void init_xml_syntax_error();
 VALUE Nokogiri_wrap_xml_syntax_error(xmlErrorPtr error);
 void Nokogiri_error_array_pusher(void * ctx, xmlErrorPtr error);
+void Nokogiri_error_silencer(void * ctx, xmlErrorPtr error);
 NORETURN(void Nokogiri_error_raise(void * ctx, xmlErrorPtr error));
 
 extern VALUE cNokogiriXmlSyntaxError;

data/lib/nokogiri.rb

@@ -61,8 +61,8 @@ require 'nokogiri/html/builder'
 #     puts link.content
 #   end
 #
-# See Nokogiri::XML::Node#css for more information about CSS searching.
-# See Nokogiri::XML::Node#xpath for more information about XPath searching.
+# See Nokogiri::XML::Searchable#css for more information about CSS searching.
+# See Nokogiri::XML::Searchable#xpath for more information about XPath searching.
 module Nokogiri
   class << self
     ###

data/lib/nokogiri/decorators/slop.rb

@@ -4,25 +4,26 @@ module Nokogiri
     # The Slop decorator implements method missing such that a methods may be
     # used instead of XPath or CSS.  See Nokogiri.Slop
     module Slop
+      # The default XPath search context for Slop
+      XPATH_PREFIX = "./"
+
       ###
       # look for node with +name+.  See Nokogiri.Slop
       def method_missing name, *args, &block
-        prefix = implied_xpath_context
-
         if args.empty?
-          list = xpath("#{prefix}#{name.to_s.sub(/^_/, '')}")
+          list = xpath("#{XPATH_PREFIX}#{name.to_s.sub(/^_/, '')}")
         elsif args.first.is_a? Hash
           hash = args.first
           if hash[:css]
             list = css("#{name}#{hash[:css]}")
           elsif hash[:xpath]
             conds = Array(hash[:xpath]).join(' and ')
-            list = xpath("#{prefix}#{name}[#{conds}]")
+            list = xpath("#{XPATH_PREFIX}#{name}[#{conds}]")
           end
         else
           CSS::Parser.without_cache do
             list = xpath(
-              *CSS.xpath_for("#{name}#{args.first}", :prefix => prefix)
+              *CSS.xpath_for("#{name}#{args.first}", :prefix => XPATH_PREFIX)
             )
           end
         end
@@ -32,9 +33,7 @@ module Nokogiri
       end
 
       def respond_to_missing? name, include_private = false
-        prefix = implied_xpath_context
-
-        list = xpath("#{prefix}#{name.to_s.sub(/^_/, '')}")
+        list = xpath("#{XPATH_PREFIX}#{name.to_s.sub(/^_/, '')}")
 
         !list.empty?
       end

data/lib/nokogiri/html/document_fragment.rb

@@ -1,8 +1,6 @@
 module Nokogiri
   module HTML
     class DocumentFragment < Nokogiri::XML::DocumentFragment
-      attr_accessor :errors
-
       ####
       # Create a Nokogiri::XML::DocumentFragment from +tags+, using +encoding+
       def self.parse tags, encoding = nil

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

@@ -2,13 +2,33 @@ module Nokogiri
   module HTML
     module SAX
       class PushParser
-        def initialize(doc = XML::SAX::Document.new, file_name = nil, encoding = 'UTF-8')
+        
+        # The Nokogiri::HTML::SAX::Document on which the PushParser will be
+        # operating
+        attr_accessor :document
+        
+        def initialize(doc = HTML::SAX::Document.new, file_name = nil, encoding = 'UTF-8')
           @document = doc
           @encoding = encoding
           @sax_parser = HTML::SAX::Parser.new(doc, @encoding)
 
           ## Create our push parser context
-          initialize_native(@sax_parser, file_name, @encoding)
+          initialize_native(@sax_parser, file_name, encoding)
+        end
+        
+        ###
+        # Write a +chunk+ of HTML to the PushParser.  Any callback methods
+        # that can be called will be called immediately.
+        def write chunk, last_chunk = false
+          native_write(chunk, last_chunk)
+        end
+        alias :<< :write
+
+        ###
+        # Finish the parsing.  This method is only necessary for
+        # Nokogiri::HTML::SAX::Document#end_document to be called.
+        def finish
+          write '', true
         end
       end
     end

data/lib/nokogiri/version.rb

@@ -1,6 +1,6 @@
 module Nokogiri
   # The version of Nokogiri you are using
-  VERSION = '1.6.5'
+  VERSION = '1.6.6.1'
 
   class VersionInfo # :nodoc:
     def jruby?

data/lib/nokogiri/xml.rb

@@ -1,6 +1,7 @@
 require 'nokogiri/xml/pp'
 require 'nokogiri/xml/parse_options'
 require 'nokogiri/xml/sax'
+require 'nokogiri/xml/searchable'
 require 'nokogiri/xml/node'
 require 'nokogiri/xml/attribute_decl'
 require 'nokogiri/xml/element_decl'

data/lib/nokogiri/xml/document.rb

@@ -5,8 +5,8 @@ module Nokogiri
     # XML documents.  The Document is created by parsing an XML document.
     # See Nokogiri::XML::Document.parse() for more information on parsing.
     #
-    # For searching a Document, see Nokogiri::XML::Node#css and
-    # Nokogiri::XML::Node#xpath
+    # For searching a Document, see Nokogiri::XML::Searchable#css and
+    # Nokogiri::XML::Searchable#xpath
     #
     class Document < Nokogiri::XML::Node
       # I'm ignoring unicode characters here.
@@ -267,8 +267,8 @@ module Nokogiri
           (string_or_io.respond_to?(:eof?) && string_or_io.eof?)
       end
 
-      def implied_xpath_context
-        "/"
+      def implied_xpath_contexts # :nodoc:
+        ["//"]
       end
 
       def inspect_attributes

data/lib/nokogiri/xml/document_fragment.rb

@@ -73,15 +73,43 @@ module Nokogiri
       end
 
       ###
-      # Search this fragment.  See Nokogiri::XML::Node#css
+      # call-seq: css *rules, [namespace-bindings, custom-pseudo-class]
+      #
+      # Search this fragment for CSS +rules+. +rules+ must be one or more CSS
+      # selectors. For example:
+      #
+      # For more information see Nokogiri::XML::Searchable#css
       def css *args
         if children.any?
-          children.css(*args)
+          children.css(*args) # 'children' is a smell here
         else
           NodeSet.new(document)
         end
       end
 
+      #
+      #  NOTE that we don't delegate #xpath to children ... another smell.
+      #  def xpath ; end
+      #
+
+      ###
+      # call-seq: search *paths, [namespace-bindings, xpath-variable-bindings, custom-handler-class]
+      #
+      # Search this fragment for +paths+. +paths+ must be one or more XPath or CSS queries.
+      #
+      # For more information see Nokogiri::XML::Searchable#search
+      def search *rules
+        rules, handler, ns, binds = extract_params(rules)
+
+        rules.inject(NodeSet.new(document)) do |set, rule|
+          set += if rule =~ Searchable::LOOKS_LIKE_XPATH
+                   xpath(*([rule, ns, handler, binds].compact))
+                 else
+                   children.css(*([rule, ns, handler].compact)) # 'children' is a smell here
+                 end
+        end
+      end
+
       alias :serialize :to_s
 
       class << self
@@ -92,6 +120,15 @@ module Nokogiri
         end
       end
 
+      # A list of Nokogiri::XML::SyntaxError found when parsing a document
+      def errors
+        document.errors
+      end
+
+      def errors= things # :nodoc:
+        document.errors = things
+      end
+
       private
 
       # fix for issue 770

data/lib/nokogiri/xml/node.rb

@@ -32,9 +32,10 @@ module Nokogiri
     # * Nokogiri::XML::Node#next
     # * Nokogiri::XML::Node#previous
     #
-    # You may search this node's subtree using Node#xpath and Node#css
+    # You may search this node's subtree using Searchable#xpath and Searchable#css
     class Node
       include Nokogiri::XML::PP::Node
+      include Nokogiri::XML::Searchable
       include Enumerable
 
       # Element node type, see Nokogiri::XML::Node#element?
@@ -90,130 +91,6 @@ module Nokogiri
         document.decorate(self)
       end
 
-      ###
-      # Search this node for +paths+.  +paths+ can be XPath or CSS, and an
-      # optional hash of namespaces may be appended.
-      # See Node#xpath and Node#css.
-      def search *paths
-        # TODO use         paths, handler, ns, binds = extract_params(paths)
-        ns = paths.last.is_a?(Hash) ? paths.pop :
-          (document.root ? document.root.namespaces : {})
-
-        prefix = "#{implied_xpath_context}/"
-
-        xpath(*(paths.map { |path|
-          path = path.to_s
-          path =~ /^(\.\/|\/|\.\.|\.$)/ ? path : CSS.xpath_for(
-            path,
-            :prefix => prefix,
-            :ns     => ns
-          )
-        }.flatten.uniq) + [ns])
-      end
-      alias :/ :search
-
-      ###
-      # call-seq: xpath *paths, [namespace-bindings, variable-bindings, custom-handler-class]
-      #
-      # Search this node for XPath +paths+. +paths+ must be one or more XPath
-      # queries.
-      #
-      #   node.xpath('.//title')
-      #
-      # A hash of namespace bindings may be appended. For example:
-      #
-      #   node.xpath('.//foo:name', {'foo' => 'http://example.org/'})
-      #   node.xpath('.//xmlns:name', node.root.namespaces)
-      #
-      # A hash of variable bindings may also be appended to the namespace bindings. For example:
-      #
-      #   node.xpath('.//address[@domestic=$value]', nil, {:value => 'Yes'})
-      #
-      # Custom XPath functions may also be defined.  To define custom
-      # functions create a class and implement the function you want
-      # to define.  The first argument to the method will be the
-      # current matching NodeSet.  Any other arguments are ones that
-      # you pass in.  Note that this class may appear anywhere in the
-      # argument list.  For example:
-      #
-      #   node.xpath('.//title[regex(., "\w+")]', Class.new {
-      #     def regex node_set, regex
-      #       node_set.find_all { |node| node['some_attribute'] =~ /#{regex}/ }
-      #     end
-      #   }.new)
-      #
-      def xpath *paths
-        return NodeSet.new(document) unless document
-
-        paths, handler, ns, binds = extract_params(paths)
-
-        sets = paths.map { |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)
-        }
-        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
-      end
-
-      ###
-      # call-seq: css *rules, [namespace-bindings, custom-pseudo-class]
-      #
-      # Search this node for CSS +rules+. +rules+ must be one or more CSS
-      # selectors. For example:
-      #
-      #   node.css('title')
-      #   node.css('body h1.bold')
-      #   node.css('div + p.green', 'div#one')
-      #
-      # A hash of namespace bindings may be appended. For example:
-      #
-      #   node.css('bike|tire', {'bike' => 'http://schwinn.com/'})
-      #
-      # Custom CSS pseudo classes may also be defined.  To define
-      # custom pseudo classes, create a class and implement the custom
-      # pseudo class you want defined.  The first argument to the
-      # method will be the current matching NodeSet.  Any other
-      # arguments are ones that you pass in.  For example:
-      #
-      #   node.css('title:regex("\w+")', Class.new {
-      #     def regex node_set, regex
-      #       node_set.find_all { |node| node['some_attribute'] =~ /#{regex}/ }
-      #     end
-      #   }.new)
-      #
-      # Note that the CSS query string is case-sensitive with regards
-      # to your document type. That is, if you're looking for "H1" in
-      # an HTML document, you'll never find anything, since HTML tags
-      # will match only lowercase CSS queries. However, "H1" might be
-      # found in an XML document, where tags names are case-sensitive
-      # (e.g., "H1" is distinct from "h1").
-      #
-      def css *rules
-        rules, handler, ns, binds = extract_params(rules)
-
-        prefix = "#{implied_xpath_context}/"
-
-        rules = rules.map { |rule|
-          CSS.xpath_for(rule, :prefix => prefix, :ns => ns)
-        }.flatten.uniq + [ns, handler, binds].compact
-
-        xpath(*rules)
-      end
-
       ###
       # Search this node's immediate children using CSS selector +selector+
       def > selector
@@ -221,33 +98,6 @@ module Nokogiri
         xpath CSS.xpath_for(selector, :prefix => "./", :ns => ns).first
       end
 
-      ###
-      # Search for the first occurrence of +path+.
-      #
-      # Returns nil if nothing is found, otherwise a Node.
-      def at path, ns = document.root ? document.root.namespaces : {}
-        search(path, ns).first
-      end
-      alias :% :at
-
-      ##
-      # Search this node for the first occurrence of XPath +paths+.
-      # Equivalent to <tt>xpath(paths).first</tt>
-      # See Node#xpath for more information.
-      #
-      def at_xpath *paths
-        xpath(*paths).first
-      end
-
-      ##
-      # Search this node for the first occurrence of CSS +rules+.
-      # Equivalent to <tt>css(rules).first</tt>
-      # See Node#css for more information.
-      #
-      def at_css *rules
-        css(*rules).first
-      end
-
       ###
       # Get the attribute value for the attribute +name+
       def [] name
@@ -327,7 +177,7 @@ 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?
-        
+
         add_sibling :next, node_or_tags
       end
 
@@ -485,7 +335,9 @@ module Nokogiri
       ###
       # Remove the attribute named +name+
       def remove_attribute name
-        attributes[name].remove if key? name
+        attr = attributes[name].remove if key? name
+        clear_xpath_context if Nokogiri.jruby?
+        attr
       end
       alias :delete :remove_attribute
 
@@ -930,25 +782,8 @@ module Nokogiri
         write_to io, options
       end
 
-      def extract_params params # :nodoc:
-        # Pop off our custom function handler if it exists
-        handler = params.find { |param|
-          ![Hash, String, Symbol].include?(param.class)
-        }
-
-        params -= [handler] if handler
-
-        hashes = []
-        while Hash === params.last || params.last.nil?
-          hashes << params.pop
-          break if params.empty?
-        end
-
-        ns, binds = hashes.reverse
-
-        ns ||= document.root ? document.root.namespaces : {}
-
-        [params, handler, ns, binds]
+      def inspect_attributes
+        [:name, :namespace, :attribute_nodes, :children]
       end
 
       def coerce data # :nodoc:
@@ -971,22 +806,17 @@ Requires a Node, NodeSet or String argument, and cannot accept a #{data.class}.
         EOERR
       end
 
-      def implied_xpath_context
-        "./"
-      end
-
-      def inspect_attributes
-        [:name, :namespace, :attribute_nodes, :children]
+      def implied_xpath_contexts # :nodoc:
+        [".//"]
       end
 
-      def add_child_node_and_reparent_attrs node
+      def add_child_node_and_reparent_attrs node # :nodoc:
         add_child_node node
         node.attribute_nodes.find_all { |a| a.name =~ /:/ }.each do |attr_node|
           attr_node.remove
           node[attr_node.name] = attr_node.value
         end
       end
-
     end
   end
 end