nokogiri 1.3.1 → 1.3.2

data/lib/nokogiri/version.rb

@@ -1,12 +1,12 @@
 module Nokogiri
   # The version of Nokogiri you are using
-  VERSION = '1.3.1'
+  VERSION = '1.3.2'
 
   # More complete version information about libxml
   VERSION_INFO = {}
   VERSION_INFO['warnings']              = []
   VERSION_INFO['nokogiri']              = VERSION
-  if defined?(LIBXML_VERSION) && ! defined?(FFI)
+  if defined?(LIBXML_VERSION)
     VERSION_INFO['libxml']              = {}
     VERSION_INFO['libxml']['binding']   = 'extension'
     VERSION_INFO['libxml']['compiled']  = LIBXML_VERSION

data/lib/nokogiri/xml.rb

@@ -24,7 +24,7 @@ module Nokogiri
   class << self
     ###
     # Parse XML.  Convenience method for Nokogiri::XML::Document.parse
-    def XML thing, url = nil, encoding = nil, options = 1, &block
+    def XML thing, url = nil, encoding = nil, options = XML::ParseOptions::DEFAULT_XML, &block
       Nokogiri::XML::Document.parse(thing, url, encoding, options, &block)
     end
   end
@@ -34,7 +34,7 @@ module Nokogiri
       ###
       # Parse an XML document using the Nokogiri::XML::Reader API.  See
       # Nokogiri::XML::Reader for mor information
-      def Reader string_or_io, url = nil, encoding = nil, options = 0
+      def Reader string_or_io, url = nil, encoding = nil, options = ParseOptions::STRICT
 
         options = Nokogiri::XML::ParseOptions.new(options) if Fixnum === options
         # Give the options to the user
@@ -48,7 +48,7 @@ module Nokogiri
 
       ###
       # Parse XML.  Convenience method for Nokogiri::XML::Document.parse
-      def parse thing, url = nil, encoding = nil, options = 1, &block
+      def parse thing, url = nil, encoding = nil, options = ParseOptions::DEFAULT_XML, &block
         Document.parse(thing, url, encoding, options, &block)
       end
 

data/lib/nokogiri/xml/document.rb

@@ -31,6 +31,14 @@ module Nokogiri
         @decorators[key] ||= []
       end
 
+      ###
+      # Validate this Document against it's DTD.  Returns a list of errors on
+      # the document or +nil+ when there is no DTD.
+      def validate
+        return nil unless internal_subset
+        internal_subset.validate self
+      end
+
       ###
       # Explore a document with shortcut methods.
       def slop!
@@ -54,6 +62,7 @@ module Nokogiri
 
       alias :to_xml :serialize
       alias :inner_html :serialize
+      alias :clone :dup
 
       # Get the hash of namespaces on the root Nokogiri::XML::Node
       def namespaces
@@ -66,7 +75,8 @@ module Nokogiri
         DocumentFragment.new(self, tags)
       end
 
-      undef_method :swap, :parent, :namespace
+      undef_method :swap, :parent, :namespace, :default_namespace=
+      undef_method :add_namespace_definition
 
       class << self
         ###
@@ -75,9 +85,9 @@ module Nokogiri
         # +url+ is resource where this document is located.  +encoding+ is the
         # encoding that should be used when processing the document. +options+
         # is a number that sets options in the parser, such as
-        # Nokogiri::XML::PARSE_RECOVER.  See the constants in
-        # Nokogiri::XML.
-        def parse string_or_io, url = nil, encoding = nil, options = 2145, &block
+        # Nokogiri::XML::ParseOptions::RECOVER.  See the constants in
+        # Nokogiri::XML::ParseOptions.
+        def parse string_or_io, url = nil, encoding = nil, options = ParseOptions::DEFAULT_XML, &block
 
           options = Nokogiri::XML::ParseOptions.new(options) if Fixnum === options
           # Give the options to the user

data/lib/nokogiri/xml/fragment_handler.rb

@@ -46,6 +46,14 @@ module Nokogiri
         @stack.last << Nokogiri::XML::Text.new(string, @document)
       end
 
+      def comment string
+        @stack.last << Nokogiri::XML::Comment.new(@document, string)
+      end
+
+      def cdata_block string
+        @stack.last << Nokogiri::XML::CDATA.new(@document, string)
+      end
+
       def end_element name
         return unless @stack.last.name == name
         @stack.pop

data/lib/nokogiri/xml/node.rb

@@ -462,7 +462,7 @@ module Nokogiri
 
       ####
       #  replace this Node with the +new_node+ in the Document.
-      def replace(new_node)
+      def replace new_node
         if new_node.is_a?(Document) || !new_node.is_a?(XML::Node)
           raise ArgumentError, <<-EOERR
 Node.replace requires a Node argument, and cannot accept a Document.
@@ -495,18 +495,6 @@ Node.replace requires a Node argument, and cannot accept a Document.
       #   end
       #
       def serialize *args, &block
-        if args.first && !args.first.is_a?(Hash)
-          $stderr.puts(<<-eowarn)
-#{self.class}#serialize(encoding, save_opts) is deprecated and will be removed in
-Nokogiri version 1.4.0 *or* after June 1 2009.
-You called serialize from here:
-
-  #{caller.first}
-
-Please change to #{self.class}#serialize(:encoding => enc, :save_with => opts)
-          eowarn
-        end
-
         options = args.first.is_a?(Hash) ? args.shift : {
           :encoding   => args[0],
           :save_with  => args[1] || SaveOptions::FORMAT
@@ -526,19 +514,6 @@ Please change to #{self.class}#serialize(:encoding => enc, :save_with => opts)
       # See Node#write_to for a list of +options+.  For formatted output,
       # use Node#to_xhtml instead.
       def to_html options = {}
-        if options.is_a?(String)
-          $stderr.puts(<<-eowarn)
-Node#to_html(encoding) is deprecated and will be removed in
-Nokogiri version 1.4.0 *or* after June 1 2009.
-You called to_html from here:
-
-  #{caller.first}
-
-Please change to Node#to_html(:encoding => #{options})
-          eowarn
-          options = { :encoding => options }
-        end
-
         # FIXME: this is a hack around broken libxml versions
         return dump_html if %w[2 6] === LIBXML_VERSION.split('.')[0..1]
 
@@ -559,21 +534,6 @@ Please change to Node#to_html(:encoding => #{options})
       def to_xml options = {}
         encoding = nil
 
-        # FIXME add a deprecation warning
-        if options.is_a? String
-          $stderr.puts(<<-eowarn)
-Node#to_xml(encoding) is deprecated and will be removed in
-Nokogiri version 1.4.0 *or* after June 1 2009.
-You called to_xml from here:
-
-  #{caller.first}
-
-Please change to Node#to_xml(:encoding => #{options})
-          eowarn
-          options = {
-            :encoding => options
-          }
-        end
         options[:save_with] ||= SaveOptions::FORMAT | SaveOptions::AS_XML
 
         serialize(options)
@@ -586,19 +546,6 @@ Please change to Node#to_xml(:encoding => #{options})
       #
       # See Node#write_to for a list of +options+
       def to_xhtml options = {}
-        if options.is_a?(String)
-          options = { :encoding => options }
-          $stderr.puts(<<-eowarn)
-Node#to_xml(encoding) is deprecated and will be removed in
-Nokogiri version 1.4.0 *or* after June 1 2009.
-You called to_xhtml from here:
-
-  #{caller.first}
-
-Please change to Node#to_xhtml(:encoding => #{options})
-          eowarn
-        end
-
         # FIXME: this is a hack around broken libxml versions
         return dump_html if %w[2 6] === LIBXML_VERSION.split('.')[0..1]
 
@@ -628,18 +575,6 @@ Please change to Node#to_xhtml(:encoding => #{options})
       #   node.write_to(io, :indent_text => '-', :indent => 2
       #
       def write_to io, *options
-        if options.length > 0 && !options.first.is_a?(Hash)
-          $stderr.puts(<<-eowarn)
-Node#write_to(io, encoding, save_options) is deprecated and will be removed in
-Nokogiri version 1.4.0 *or* after June 1 2009.
-You called write_to from here:
-
-  #{caller.first}
-
-Please change to: Node#write_to(io, :encoding => e, :save_options => opts)
-          eowarn
-        end
-
         options       = options.first.is_a?(Hash) ? options.shift : {}
         encoding      = options[:encoding] || options[0]
         save_options  = options[:save_with] || options[1] || SaveOptions::FORMAT
@@ -658,19 +593,6 @@ Please change to: Node#write_to(io, :encoding => e, :save_options => opts)
       #
       # See Node#write_to for a list of +options+
       def write_html_to io, options = {}
-        if options.is_a?(String)
-          $stderr.puts(<<-eowarn)
-Node#write_html_to(io, encoding) is deprecated and will be removed in
-Nokogiri version 1.4.0 *or* after June 1 2009.
-You called write_html_to from here:
-
-  #{caller.first}
-
-Please change to Node#write_html_to(io, :encoding => #{options})
-          eowarn
-          options = { :encoding => options }
-        end
-
         # FIXME: this is a hack around broken libxml versions
         return (io << dump_html) if %w[2 6] === LIBXML_VERSION.split('.')[0..1]
 
@@ -686,19 +608,6 @@ Please change to Node#write_html_to(io, :encoding => #{options})
       #
       # See Node#write_to for a list of +options+
       def write_xhtml_to io, options = {}
-        if options.is_a?(String)
-          $stderr.puts(<<-eowarn)
-Node#write_xhtml_to(io, encoding) is deprecated and will be removed in
-Nokogiri version 1.4.0 *or* after June 1 2009.
-You called write_xhtml_to from here:
-
-  #{caller.first}
-
-Please change to Node#write_xhtml_to(io, :encoding => #{options})
-          eowarn
-          options = { :encoding => options }
-        end
-
         # FIXME: this is a hack around broken libxml versions
         return (io << dump_html) if %w[2 6] === LIBXML_VERSION.split('.')[0..1]
 
@@ -716,18 +625,6 @@ Please change to Node#write_xhtml_to(io, :encoding => #{options})
       #
       # See Node#write_to for a list of options
       def write_xml_to io, options = {}
-        if options.is_a?(String)
-          $stderr.puts(<<-eowarn)
-Node#write_xml_to(io, encoding) is deprecated and will be removed in
-Nokogiri version 1.4.0 *or* after June 1 2009.
-You called write_xml_to from here:
-
-  #{caller.first}
-
-Please change to Node#write_xml_to(io, :encoding => #{options})
-          eowarn
-          options = { :encoding => options }
-        end
         options[:save_with] ||= SaveOptions::FORMAT | SaveOptions::AS_XML
         write_to io, options
       end

data/lib/nokogiri/xml/node_set.rb

@@ -68,19 +68,59 @@ module Nokogiri
       # For more information see Nokogiri::XML::Node#css and
       # Nokogiri::XML::Node#xpath
       def search *paths
-        ns = paths.last.is_a?(Hash) ? paths.pop : document.root.namespaces
+        ns = paths.last.is_a?(Hash) ? paths.pop :
+          (document.root ? document.root.namespaces : {})
+
         sub_set = NodeSet.new(document)
-        document.decorate(sub_set)
         each do |node|
-          node.search(*(paths + [ns])).each do |sub_node|
-            sub_set << sub_node
+          paths.each do |path|
+            sub_set +=
+              send(path =~ /^(\.\/|\/)/ ? :xpath : :css, *(paths + [ns]))
           end
         end
+        document.decorate(sub_set)
         sub_set
       end
       alias :/ :search
-      alias :xpath :search
-      alias :css :search
+
+      ###
+      # Search this NodeSet for css +paths+
+      #
+      # For more information see Nokogiri::XML::Node#css
+      def css *paths
+        ns = paths.last.is_a?(Hash) ? paths.pop :
+          (document.root ? document.root.namespaces : {})
+
+        sub_set = NodeSet.new(document)
+
+        xpaths = paths.map { |rule|
+          [
+            CSS.xpath_for(rule.to_s, :prefix => ".//", :ns => ns),
+            CSS.xpath_for(rule.to_s, :prefix => "self::", :ns => ns)
+          ].join(' | ')
+        }
+        each do |node|
+          sub_set += node.xpath(*(xpaths + [ns]))
+        end
+        document.decorate(sub_set)
+        sub_set
+      end
+
+      ###
+      # Search this NodeSet for XPath +paths+
+      #
+      # For more information see Nokogiri::XML::Node#xpath
+      def xpath *paths
+        ns = paths.last.is_a?(Hash) ? paths.pop :
+          (document.root ? document.root.namespaces : {})
+
+        sub_set = NodeSet.new(document)
+        each do |node|
+          sub_set += node.xpath(*(paths + [ns]))
+        end
+        document.decorate(sub_set)
+        sub_set
+      end
 
       ###
       # If path is a string, search this document for +path+ returning the

data/lib/nokogiri/xml/parse_options.rb

@@ -38,13 +38,18 @@ module Nokogiri
       # do not generate XINCLUDE START/END nodes
       NOXINCNODE  = 1 << 15
 
+      # the default options used for parsing XML documents
+      DEFAULT_XML  = RECOVER
+      # the default options used for parsing HTML documents
+      DEFAULT_HTML = RECOVER | NOERROR | NOWARNING | NONET
+
       attr_accessor :options
-      def initialize options = 0
+      def initialize options = STRICT
         @options = options
       end
 
       constants.each do |constant|
-        next if constant == 'STRICT'
+        next if constant.to_sym == :STRICT
         class_eval %{
           def #{constant.downcase}
             @options |= #{constant}

data/lib/nokogiri/xml/relax_ng.rb

@@ -18,8 +18,8 @@ module Nokogiri
     # Validate an XML document against a RelaxNG schema.  Loop over the errors
     # that are returned and print them out:
     #
-    #   schema  = Nokogiri::XML::RelaxNG(File.read(ADDRESS_SCHEMA_FILE))
-    #   doc     = Nokogiri::XML(File.read(ADDRESS_XML_FILE))
+    #   schema  = Nokogiri::XML::RelaxNG(File.open(ADDRESS_SCHEMA_FILE))
+    #   doc     = Nokogiri::XML(File.open(ADDRESS_XML_FILE))
     #
     #   schema.validate(doc).each do |error|
     #     puts error.message

data/lib/nokogiri/xml/sax.rb

@@ -1,3 +1,4 @@
 require 'nokogiri/xml/sax/document'
+require 'nokogiri/xml/sax/legacy_handlers'
 require 'nokogiri/xml/sax/parser'
 require 'nokogiri/xml/sax/push_parser'

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

@@ -93,11 +93,11 @@ module Nokogiri
         ###
         # Called at the beginning of an element
         # +name+ is the element name
-        # +attrs+ is a hash of attributes
+        # +attrs+ is a list of attributes
         # +prefix+ is the namespace prefix for the element
         # +uri+ is the associated namespace URI
-        # +namespaces+ is a hash of namespace prefix:urls associated with the element
-        def start_element_ns(name, attrs = {}, prefix = nil, uri = nil, namespaces = {})
+        # +ns+ is a hash of namespace prefix:urls associated with the element
+        def start_element_namespace name, attrs = {}, prefix = nil, uri = nil, ns = {}
         end
 
         ###
@@ -105,7 +105,7 @@ module Nokogiri
         # +name+ is the element's name
         # +prefix+ is the namespace prefix associated with the element
         # +uri+ is the associated namespace URI
-        def end_element_ns(name, prefix = nil, uri = nil)
+        def end_element_namespace name, prefix = nil, uri = nil
         end
 
         ###

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

@@ -0,0 +1,65 @@
+module Nokogiri
+  module XML
+    module SAX
+      # :stopdoc:
+      module LegacyHandlers
+        def start_element_namespace name,
+                                    attrs   = [],
+                                    prefix  = nil,
+                                    uri     = nil,
+                                    ns      = []
+
+          ##
+          # Deal with legacy interface
+          if @document.respond_to? :start_element_ns
+            unless @warned
+              warn <<-eowarn
+Nokogiri::XML::SAX::Document#start_element_ns and end_element_ns are deprecated,
+please change to start_element_namespace.  start_element_ns will be removed by
+version 1.4.0 or by August 1st, whichever comes first.
+              eowarn
+              @warned = true
+            end
+            attr_hash = {}
+            attrs.each do |attr|
+              attr_hash[attr.localname] = attr.value
+            end
+            ns_hash = Hash[*ns.flatten]
+            @document.start_element_ns name, attr_hash, prefix, uri, ns_hash
+          end
+
+          ###
+          # Deal with SAX v1 interface
+          name = [prefix, name].compact.join(':')
+          attributes = ns.map { |ns_prefix,ns_uri|
+            [['xmlns', ns_prefix].compact.join(':'), ns_uri]
+          } + attrs.map { |attr|
+            [[attr.prefix, attr.localname].compact.join(':'), attr.value]
+          }.flatten
+          @document.start_element name, attributes
+        end
+
+        def end_element_namespace name, prefix = nil, uri = nil
+          ##
+          # Deal with legacy interface
+          if @document.respond_to? :end_element_ns
+            unless @warned
+              warn <<-eowarn
+Nokogiri::XML::SAX::Document#start_element_ns and end_element_ns are deprecated,
+please change to start_element_namespace.  start_element_ns will be removed by
+version 1.4.0 or by August 1st, whichever comes first.
+              eowarn
+              @warned = true
+            end
+            @document.end_element_ns name, prefix, uri
+          end
+
+          ###
+          # Deal with SAX v1 interface
+          @document.end_element [prefix, name].compact.join(':')
+        end
+      end
+      # :startdoc:
+    end
+  end
+end