vcdom 0.3.0 → 0.3.1

data/README.rdoc

@@ -0,0 +1,33 @@
+= VCDOM [RubyGem]
+
+This project makes a implementation of W3C DOM working on Ruby.
+Now, it's under construction...
+
+== Example
+
+following example shows you how to parse XML string and how to serialize DOM objects:
+
+  require "vcdom/xml_ls"
+  
+  @ls_impl = VCDOM::XMLLS
+  # get parser
+  parser = @ls_impl.create_ls_parser( :mode_asynchronous, nil )
+  
+  # set input xml string
+  input  = @ls_impl.create_ls_input()
+  input.string_data = "<empty-element/>"
+  # do parse
+  doc = parser.parse( input )
+  doc.document_element.tag_name
+      #=> "empty-element"
+  
+  # parse other xml string
+  input.string_data = "<test>aaaa</test>"
+  doc = parser.parse( input )
+  doc.document_element.first_child.node_value
+      #=> "aaaa"
+  
+  # serialize
+  serializer = @ls_impl.create_ls_serializer()
+  serializer.write_to_string( doc ) 
+      #=> "<test>aaa</test>"

data/lib/vcdom/attr.rb

@@ -8,7 +8,7 @@ module VCDOM
       
       include Parent
       
-      def initialize( doc, name )
+      def initialize( doc, name ) # :nodoc:
         initialize_parent()
         super( doc )
         @owner_element  = nil
@@ -22,7 +22,7 @@ module VCDOM
       def owner_element
         @owner_element
       end
-      def _set_owner_element( elem )
+      def _set_owner_element( elem ) # :nodoc:
         @owner_element = elem
       end
       
@@ -51,7 +51,7 @@ module VCDOM
       end
       
       def append_child( new_child )
-        # ノードのタイプチェックなど
+        # check the arg type
         if not new_child.is_a? Node then
           raise ArgumentError.new( "the argument [#{new_child.inspect}] is not an object of the expected class." )
         end

data/lib/vcdom/character_data.rb

@@ -13,7 +13,14 @@ module VCDOM
       def data
         @data.to_s
       end
-      alias :node_value :data
+      def data=( val )
+        @data = val.intern
+      end
+      alias :node_value    :data
+      alias :node_value=   :data=
+      alias :text_content  :data
+      alias :text_content= :data=
+      
       def length
         @data.to_s.length
       end

data/lib/vcdom/document.rb

@@ -13,8 +13,7 @@ module VCDOM
       
       include Parent
       
-      
-      def initialize()
+      def initialize() # :nodoc:
         initialize_parent()
         super( nil )
       end
@@ -23,30 +22,50 @@ module VCDOM
         DOCUMENT_NODE
       end
       
+      def document_element
+        @document_element
+      end
+      
+      # always +nil+
+      def text_content
+        nil
+      end
+      
+      # no effect
+      def text_content=( val )
+      end
+      
       def append_child( new_child )
-        # ノードのタイプチェックなど
+        # Check the arg type
         if not new_child.is_a? Node then
           raise ArgumentError.new( "the argument [#{new_child.inspect}] is not an object of the expected class." )
         end
         # Element (maximum of one), ProcessingInstruction, Comment, DocumentType (maximum of one) 
         case new_child.node_type
           when ELEMENT_NODE then
-            # 既に追加されてるかどうかのチェック
+            # is there document_element already?
             if @document_element.nil? then
               @document_element = new_child
             else
               raise "HIERARCHY_REQUEST_ERR"
             end
           when DOCUMENT_TYPE_NODE then
-            # 既に追加されてるかどうかのチェック
+            # is there document_type already?
           when PROCESSING_INSTRUCTION_NODE, COMMENT_NODE then
             # OK
           else
             # ERROR
-            raise "ERROR"
+            raise "ERROR : new_child.node_type = #{new_child.node_type}"
         end
         _append_child( new_child )
       end
+      def remove_child( old_child )
+        old_child = super( old_child )
+        if old_child.equal? @document_element then
+          @document_element = nil
+        end
+        return old_child
+      end
       
       def create_element( tag_name )
         elem = nil
@@ -113,13 +132,10 @@ module VCDOM
       end
       
       def create_text_node( data )
-        node = nil
-        if data.is_a? String then
-          node = Text._new( self, data.intern )
-        else
+        if not data.is_a? String then
           raise ArgumentError.new( "the argument [#{data.inspect}] is not an object of the expected class. the class : #{data.class}" )
         end
-        node
+        node = Text._new( self, data.intern )
       end
       
     end

data/lib/vcdom/element.rb

@@ -11,7 +11,7 @@ module VCDOM
       include Parent
       include Child
       
-      def initialize( doc, tag_name )
+      def initialize( doc, tag_name ) # :nodoc:
         initialize_parent()
         super( doc )
         @local_name = tag_name
@@ -84,6 +84,18 @@ module VCDOM
         self.set_attribute_node( attr )
         nil
       end
+      # Retrieves an attribute value by name.
+      def get_attribute( name )
+        @attr_nodes.each do |attr|
+          if attr.name == name then
+            #//////////////////////////
+            # 変更する
+            #//////////////////////////
+            return attr.value
+          end
+        end
+        return ""
+      end
       
       # setAttributeNodeNS introduced in DOM Level 2
       # Adds a new attribute. If an attribute with that local name and that namespace URI is already present in the element, it is replaced by the new one. Replacing an attribute node by itself has no effect.
@@ -328,5 +340,79 @@ module VCDOM
         end # end Element.normalizeNamespaces
       end
       
+      # The first child node of that element which is of nodeType +ELEMENT_NODE+, as an Element object.
+      # If the element on which this attribute is accessed does not have any child nodes, 
+      # or if none of those child nodes are element nodes, then this attribute return +nil+. 
+      # (defiend at W3C Element Traversal Specification)
+      def first_element_child
+        n = self.first_child
+        while n do
+          if n.node_type == ELEMENT_NODE then
+            break
+          end
+          n = n.next_sibling
+        end
+        return n
+      end
+      
+      # The last child node of that element which is of nodeType +ELEMENT_NODE+, as an Element object.
+      # If the element on which this attribute is accessed does not have any child nodes, 
+      # or if none of those child nodes are element nodes, then this attribute return +nil+. 
+      # (defiend at W3C Element Traversal Specification)
+      def last_element_child
+        n = self.last_child
+        while n do
+          if n.node_type == ELEMENT_NODE then
+            break
+          end
+          n = n.previous_sibling
+        end
+        return n
+      end
+      
+      # The sibling node of that element which most immediately precedes that element in document order, 
+      # and which is of nodeType +ELEMENT_NODE+, as an Element object. 
+      # If the element on which this attribute is accessed does not have any preceding sibling nodes, 
+      # or if none of those preceding sibling nodes are element nodes, then this attribute must return +nil+. 
+      # (defiend at W3C Element Traversal Specification)
+      def previous_element_sibling
+        n = self.previous_sibling
+        while n do
+          if n.node_type == ELEMENT_NODE then
+            break
+          end
+          n = n.previous_sibling
+        end
+        return n
+      end
+      
+      # The sibling node of that element which most immediately follows that element in document order, 
+      # and which is of nodeType +ELEMENT_NODE+, as an Element object. 
+      # If the element on which this attribute is accessed does not have any following sibling nodes, 
+      # or if none of those following sibling nodes are element nodes, then this attribute must return +nil+. 
+      # (defiend at W3C Element Traversal Specification)
+      def next_element_sibling
+        n = self.next_sibling
+        while n do
+          if n.node_type == ELEMENT_NODE then
+            break
+          end
+          n = n.next_sibling
+        end
+        return n
+      end
+      
+      # The current number of child nodes of that element which are of nodeType +ELEMENT_NODE+. 
+      # This value is not stored, but calculated when you access this attribute. 
+      def child_element_count
+        num = 0
+        self.each_child_node do |n|
+          if n.node_type == ELEMENT_NODE then
+            num += 1
+          end
+        end
+        return num
+      end
+      
     end
 end

data/lib/vcdom/node.rb

@@ -21,7 +21,7 @@ module VCDOM
         private :new
       end
       
-      def initialize( owner_document )
+      def initialize( owner_document ) # :nodoc:
         @owner_document = owner_document
       end
       
@@ -58,7 +58,29 @@ module VCDOM
         nil
       end
       
+      # Concatenation of the textContent attribute value of every child node, 
+      # excluding +COMMENT_NODE+ and +PROCESSING_INSTRUCTION_NODE+ nodes. 
+      # This is the empty string if the node has no children.
+      def text_content
+        str = String.new()
+        self.each_child_node do |n|
+          if n.node_type != COMMENT_NODE and n.node_type != PROCESSING_INSTRUCTION_NODE then
+            str << n.text_content
+          end
+        end
+        return str
+      end
       
+      # Any possible children this node have are removed and, if the new value is not empty +String+ object or +nil+, 
+      # replaced by a single Text node containing the new value string. 
+      def text_content=( val )
+        n = self.first_child
+        while n do
+          self.remove_child( n )
+          n = self.first_child
+        end
+        self.append_child( self.owner_document.create_text_node( val ) )
+      end
       
       def lookup_prefix( namespace_uri )
         if namespace_uri.nil? then #(namespaceURI has no value, i.e. namespaceURI is null or empty string) {
@@ -100,6 +122,12 @@ module VCDOM
             return nil
         end
       end
+      
+      # Always +false+.
+      def has_child_nodes()
+        return false
+      end
+      
       def lookup_namespace_prefix( namespace_uri, original_element )
         if not self.namespace_uri.nil? and self.namespace_uri == namespace_uri and 
               not self.prefix.nil? and original_element.lookup_namespace_uri( self.prefix ) == namespace_uri then

data/lib/vcdom/parent.rb

@@ -70,6 +70,15 @@ module VCDOM
         @child_nodes.last_child
       end
       
+      # Returns whether this node has any children. 
+      def has_child_nodes()
+        if @child_nodes.first_child.nil? then
+          return false
+        else
+          return true
+        end
+      end
+      
       def _append_child( new_child )
         #@child_nodes << new_child
         if not new_child.parent_node.nil? then

data/lib/vcdom/{xml_parser.rb → xml_ls/xml_parser.rb}

@@ -2,110 +2,129 @@
 
 require "vcdom/document"
 
-module VCDOM
-    class XMLInput
-      def string_data
-        @string_data
-      end
-      def string_data=( str_data )
-        @string_data = str_data
-      end
+module VCDOM::XMLLS
+  
+  class XMLInput
+    def string_data
+      @string_data
     end
-    class XMLContentHandler
-      def initialize( document_class )
-        @document_class = document_class
-        @ns_stack = Array.new()
-      end
-      def ns_stack
-        @ns_stack
-      end
-      def start_document
-        @doc = @document_class._new()
-        @cur = @doc
-      end
-      def end_document
-        return @doc
-      end
-      def on_stag( name, ns_uri )
-        @cur = @cur.append_child( @doc.create_element_ns( ns_uri, name ) )
+    def string_data=( str_data )
+      @string_data = str_data
+    end
+  end
+  
+  class XMLContentHandler # :nodoc:
+    def initialize( document_class )
+      @document_class = document_class
+      @ns_stack = Array.new()
+    end
+    def ns_stack
+      @ns_stack
+    end
+    def start_document
+      @doc = @document_class._new()
+      @cur = @doc
+    end
+    def end_document
+      return @doc
+    end
+    def on_stag( name, ns_uri )
+      @cur = @cur.append_child( @doc.create_element_ns( ns_uri, name ) )
+    end
+    def on_etag( name )
+      if @cur.node_name != name then
+        raise "[ERROR] @cur.node_name : #{@cur.node_name}, name : #{name}"
       end
-      def on_etag( name )
-        if @cur.node_name != name then
-          raise "[ERROR] @cur.node_name : #{@cur.node_name}, name : #{name}"
+      @cur = @cur.parent_node
+    end
+    def on_eetag( name, ns_uri )
+      @cur = @cur.append_child( @doc.create_element_ns( ns_uri, name ) )
+    end
+    def on_end_eetag()
+      @cur = @cur.parent_node
+    end
+    def on_begin_attr( name, ns_uri )
+      attr = @doc.create_attribute_ns( ns_uri, name )
+      @cur.set_attribute_node_ns( attr )
+      @cur = attr
+    end
+    def on_end_attr()
+      @cur = @cur.owner_element
+    end
+    def on_chardata( char_data )
+      if @cur.node_type == VCDOM::Node::DOCUMENT_NODE then
+        if char_data !~ /\A[\x20\x09\x0D\x0A]*\Z/ then
+          raise "ERROR : " + char_data.inspect
         end
-        @cur = @cur.parent_node
-      end
-      def on_eetag( name, ns_uri )
-        @cur = @cur.append_child( @doc.create_element_ns( ns_uri, name ) )
-      end
-      def on_end_eetag()
-        @cur = @cur.parent_node
-      end
-      def on_begin_attr( name, ns_uri )
-        attr = @doc.create_attribute_ns( ns_uri, name )
-        @cur.set_attribute_node_ns( attr )
-        @cur = attr
-      end
-      def on_end_attr()
-        @cur = @cur.owner_element
-      end
-      def on_chardata( char_data )
+      else
         @cur.append_child( @doc.create_text_node( char_data ) )
       end
     end
-    class XMLParser
-      
-      def parse( input )
-        source = input.string_data
-        if source.class == String then
-          _parse_xml_str( source, XMLContentHandler.new( VCDOM::Document ) )
-        else
-          raise ArgumentTypeError.new()
+  end
+  
+  class XMLParser
+    
+    def parse( input )
+      source = input.string_data
+      if source.class == String then
+        # Ruby 1.8 の場合, String を 1 文字ずつに分ける
+        if RUBY_VERSION < "1.9"
+          source = source.split //u
+          class << source
+            def []( *args )
+              res = super( *args )
+              args.length == 2 ? res.join : res
+            end
+          end
         end
+        _parse_xml_str( source, XMLContentHandler.new( VCDOM::Document ) )
+      else
+        raise ArgumentTypeError.new()
       end
-      
-      def _parse_xml_str( xml_str, content_handler )
-        content_handler.start_document()
-        i = 0
-        loop do
-          if xml_str[i].nil? then
-            break
-          elsif xml_str[i] == "<" then
+    end
+    
+    def _parse_xml_str( xml_str, content_handler )
+      content_handler.start_document()
+      i = 0
+      loop do
+        if xml_str[i].nil? then
+          break
+        elsif xml_str[i] == "<" then
+          i += 1
+          if xml_str[i] == "/" then
+            # ETag ::= '</' Name S? '>'
+            i = _parse_end_tag( xml_str, i+1, content_handler )
+          elsif xml_str[i] == "?" then
+            # PI
+          elsif xml_str[i] == "!" then
             i += 1
-            if xml_str[i] == "/" then
-              # ETag ::= '</' Name S? '>'
-              i = _parse_end_tag( xml_str, i+1, content_handler )
-            elsif xml_str[i] == "?" then
-              # PI
-            elsif xml_str[i] == "!" then
-              i += 1
-              if xml_str[i,2] == "--" then
-                # Comment ::= '<!--' ((Char - '-') | ('-' (Char - '-')))* '-->'
-                i = _parse_comment( xml_str, i+2, content_handler )
-              elsif xml_str[i,7] == "[CDATA[" then
-                # CDATA
-                i = _parse_cdata( xml_str, i+7, content_handler )
-              else
-                $stdout << "ERROR" << "\n"
-              end
+            if xml_str[i,2] == "--" then
+              # Comment ::= '<!--' ((Char - '-') | ('-' (Char - '-')))* '-->'
+              i = _parse_comment( xml_str, i+2, content_handler )
+            elsif xml_str[i,7] == "[CDATA[" then
+              # CDATA
+              i = _parse_cdata( xml_str, i+7, content_handler )
             else
-              i = _parse_stag_or_eetag( xml_str, i, content_handler )
+              $stdout << "ERROR" << "\n"
             end
-          elsif xml_str[i] == "&" and not _is_char_ref_or_predef_entity_ref?( xml_str, i ) then
-            raise "NOT SUPPORT"
-            # Reference ::= EntityRef | CharRef
-            #   EntityRef ::= '&' Name ';'
-            #   CharRef   ::= '&#' [0-9]+ ';' | '&#x' [0-9a-fA-F]+ ';'
           else
-            # CharData ::= [^<&]* - ([^<&]* ']]>' [^<&]*)
-            i = _parse_chardata_or_entity_reference( xml_str, i, content_handler )
+            i = _parse_stag_or_eetag( xml_str, i, content_handler )
           end
+        elsif xml_str[i] == "&" and not _is_char_ref_or_predef_entity_ref?( xml_str, i ) then
+          raise "NOT SUPPORT"
+          # Reference ::= EntityRef | CharRef
+          #   EntityRef ::= '&' Name ';'
+          #   CharRef   ::= '&#' [0-9]+ ';' | '&#x' [0-9a-fA-F]+ ';'
+        else
+          # CharData ::= [^<&]* - ([^<&]* ']]>' [^<&]*)
+          i = _parse_chardata_or_entity_reference( xml_str, i, content_handler )
         end
-        return content_handler.end_document()
       end
-      private :_parse_xml_str
-      
-      def _parse_stag_or_eetag( str, i, content_handler )
+      return content_handler.end_document()
+    end
+    private :_parse_xml_str
+    
+    def _parse_stag_or_eetag( str, i, content_handler )
         #@listener.on_begin_stag_or_eetag()
         # 要素名の取得
         elem_name = String.new()
@@ -253,6 +272,7 @@ module VCDOM
         end
         return i
       end
+      private :_parse_stag_or_eetag
       
       def _parse_end_tag( str, i, content_handler )
         # 要素名の取得
@@ -280,6 +300,7 @@ module VCDOM
         content_handler.ns_stack.pop()
         return i
       end
+      private :_parse_end_tag
       
       def _is_char_ref_or_predef_entity_ref?( str, i )
         if str[i] == "&" then
@@ -293,6 +314,7 @@ module VCDOM
           end
         end
       end
+      private :_is_char_ref_or_predef_entity_ref?
       
       def _parse_chardata_or_entity_reference( str, i, content_handler )
         # CharData ::= [^<&]* - ([^<&]* ']]>' [^<&]*)
@@ -304,16 +326,21 @@ module VCDOM
           #  break
           elsif str[i] == "&" then
             if str[i+1] != "#" then
-              if str[i+2,3] == "lt;" then
+              if str[i+1,3] == "lt;" then
                 chardata << "<"
-              elsif str[i+2,3] == "gt;" then
+                i += 4
+              elsif str[i+1,3] == "gt;" then
                 chardata << ">"
-              elsif str[i+2,5] == "quot;" then
+                i += 4
+              elsif str[i+1,5] == "quot;" then
                 chardata << "\""
-              elsif str[i+2,5] == "apos;" then
+                i += 6
+              elsif str[i+1,5] == "apos;" then
                 chardata << "'"
-              elsif str[i+2,4] == "amp;" then
+                i += 6
+              elsif str[i+1,4] == "amp;" then
                 chardata << "&"
+                i += 5
               else
                 # entity reference
                 break
@@ -352,6 +379,7 @@ module VCDOM
         content_handler.on_chardata( chardata )
         return i
       end
+      private :_parse_chardata_or_entity_reference
       
       def _skip_white_spaces( str, i )
         loop do
@@ -363,6 +391,7 @@ module VCDOM
         end
         return i
       end
+      private :_skip_white_spaces
       
     end
 end

data/lib/vcdom/{xml_serializer.rb → xml_ls/xml_serializer.rb}

@@ -2,7 +2,7 @@
 
 require "vcdom/node"
 
-module VCDOM
+module VCDOM::XMLLS
     class XMLSerializer
       
       def write_to_string( node )

data/lib/vcdom/xml_ls.rb

@@ -0,0 +1,42 @@
+# coding : utf-8
+
+require "vcdom/xml_ls/xml_parser"
+require "vcdom/xml_ls/xml_serializer"
+
+module VCDOM
+  module XMLLS
+    
+    extend self
+    
+    MODE_SYNCHRONOUS  = :mode_synchronous
+    MODE_ASYNCHRONOUS = :mode_asynchronous
+    
+    def create_ls_parser( mode, schema_type )
+      case mode
+      when MODE_ASYNCHRONOUS
+        # OK : do nothing
+      when MODE_SYNCHRONOUS
+        raise "SORRY... arg +mode+ is not supported. must be +MODE_ASYNCHRONOUS+"
+      else
+        raise ArgumentError.new( "invalid mode [" + mode.inspect + "]" )
+      end
+      if not schema_type.nil? then
+        raise "SORRY... arg +schema_type+ is not still supported. must be +nil+"
+      end
+      XMLParser.new()
+    end
+    
+    def create_ls_serializer()
+      XMLSerializer.new()
+    end
+    
+    def create_ls_input()
+      XMLInput.new()
+    end
+    
+    def create_ls_output()
+      XMLOutput.new()
+    end
+    
+  end
+end