nokogiri 1.0.0 → 1.6.8.1

data/lib/nokogiri/xml/document.rb

@@ -1,29 +1,284 @@
 module Nokogiri
   module XML
-    class Document < Node
-      def decorators
-        @decorators ||= Hash.new { |h,k| h[k] = [] }
+    ##
+    # Nokogiri::XML::Document is the main entry point for dealing with
+    # 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::Searchable#css and
+    # Nokogiri::XML::Searchable#xpath
+    #
+    class Document < Nokogiri::XML::Node
+      # I'm ignoring unicode characters here.
+      # See http://www.w3.org/TR/REC-xml-names/#ns-decl for more details.
+      NCNAME_START_CHAR = "A-Za-z_"
+      NCNAME_CHAR       = NCNAME_START_CHAR + "\\-.0-9"
+      NCNAME_RE         = /^xmlns(:[#{NCNAME_START_CHAR}][#{NCNAME_CHAR}]*)?$/
+
+      ##
+      # Parse an XML file.
+      #
+      # +string_or_io+ may be a String, or any object that responds to
+      # _read_ and _close_ such as an IO, or StringIO.
+      #
+      # +url+ (optional) is the URI where this document is located.
+      #
+      # +encoding+ (optional) is the encoding that should be used when processing
+      # the document.
+      #
+      # +options+ (optional) is a configuration object that sets options during
+      # parsing, such as Nokogiri::XML::ParseOptions::RECOVER. See the
+      # Nokogiri::XML::ParseOptions for more information.
+      #
+      # +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 }
+      #
+      # Nokogiri.XML() is a convenience method which will call this method.
+      #
+      def self.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
+        yield options if block_given?
+
+        if empty_doc?(string_or_io)
+          if options.strict?
+            raise Nokogiri::XML::SyntaxError.new("Empty document")
+          else
+            return encoding ? new.tap { |i| i.encoding = encoding } : new
+          end
+        end
+
+        doc = if string_or_io.respond_to?(:read)
+          url ||= string_or_io.respond_to?(:path) ? string_or_io.path : nil
+          read_io(string_or_io, url, encoding, options.to_i)
+        else
+          # read_memory pukes on empty docs
+          read_memory(string_or_io, url, encoding, options.to_i)
+        end
+
+        # do xinclude processing
+        doc.do_xinclude(options) if options.xinclude?
+
+        return doc
       end
 
+      # A list of Nokogiri::XML::SyntaxError found when parsing a document
+      attr_accessor :errors
+
+      def initialize *args # :nodoc:
+        @errors     = []
+        @decorators = nil
+      end
+
+      ##
+      # Create an element with +name+, and optionally setting the content and attributes.
+      #
+      #   doc.create_element "div" # <div></div>
+      #   doc.create_element "div", :class => "container" # <div class='container'></div>
+      #   doc.create_element "div", "contents" # <div>contents</div>
+      #   doc.create_element "div", "contents", :class => "container" # <div class='container'>contents</div>
+      #   doc.create_element "div" { |node| node['class'] = "container" } # <div class='container'></div>
+      #
+      def create_element name, *args, &block
+        elm = Nokogiri::XML::Element.new(name, self, &block)
+        args.each do |arg|
+          case arg
+          when Hash
+            arg.each { |k,v|
+              key = k.to_s
+              if key =~ NCNAME_RE
+                ns_name = key.split(":", 2)[1]
+                elm.add_namespace_definition ns_name, v
+              else
+                elm[k.to_s] = v.to_s
+              end
+            }
+          else
+            elm.content = arg
+          end
+        end
+        if ns = elm.namespace_definitions.find { |n| n.prefix.nil? or n.prefix == '' }
+          elm.namespace = ns
+        end
+        elm
+      end
+
+      # Create a Text Node with +string+
+      def create_text_node string, &block
+        Nokogiri::XML::Text.new string.to_s, self, &block
+      end
+
+      # Create a CDATA Node containing +string+
+      def create_cdata string, &block
+        Nokogiri::XML::CDATA.new self, string.to_s, &block
+      end
+
+      # Create a Comment Node containing +string+
+      def create_comment string, &block
+        Nokogiri::XML::Comment.new self, string.to_s, &block
+      end
+
+      # The name of this document.  Always returns "document"
       def name
         'document'
       end
 
-      ###
-      # Apply any decorators to +node+
-      def decorate(node)
-        key = node.class.name.split('::').last.downcase
-        decorators[key].each do |klass|
-          node.extend(klass)
+      # A reference to +self+
+      def document
+        self
+      end
+
+      ##
+      # Recursively get all namespaces from this node and its subtree and
+      # return them as a hash.
+      #
+      # For example, given this document:
+      #
+      #   <root xmlns:foo="bar">
+      #     <bar xmlns:hello="world" />
+      #   </root>
+      #
+      # This method will return:
+      #
+      #   { 'xmlns:foo' => 'bar', 'xmlns:hello' => 'world' }
+      #
+      # WARNING: this method will clobber duplicate names in the keys.
+      # For example, given this document:
+      #
+      #   <root xmlns:foo="bar">
+      #     <bar xmlns:foo="baz" />
+      #   </root>
+      #
+      # The hash returned will look like this: { 'xmlns:foo' => 'bar' }
+      #
+      # Non-prefixed default namespaces (as in "xmlns=") are not included
+      # in the hash.
+      #
+      # Note that this method does an xpath lookup for nodes with
+      # namespaces, and as a result the order may be dependent on the
+      # implementation of the underlying XML library.
+      #
+      def collect_namespaces
+        xpath("//namespace::*").inject({}) do |hash, ns|
+          hash[["xmlns",ns.prefix].compact.join(":")] = ns.href if ns.prefix != "xml"
+          hash
         end
       end
 
-      def to_xml
-        serialize
+      # Get the list of decorators given +key+
+      def decorators key
+        @decorators ||= Hash.new
+        @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. See Nokogiri::Slop for details.
+      #
+      # Note that any nodes that have been instantiated before #slop!
+      # is called will not be decorated with sloppy behavior. So, if you're in
+      # irb, the preferred idiom is:
+      #
+      #   irb> doc = Nokogiri::Slop my_markup
+      #
+      # and not
+      #
+      #   irb> doc = Nokogiri::HTML my_markup
+      #   ... followed by irb's implicit inspect (and therefore instantiation of every node) ...
+      #   irb> doc.slop!
+      #   ... which does absolutely nothing.
+      #
+      def slop!
+        unless decorators(XML::Node).include? Nokogiri::Decorators::Slop
+          decorators(XML::Node) << Nokogiri::Decorators::Slop
+          decorate!
+        end
+
+        self
+      end
+
+      ##
+      # Apply any decorators to +node+
+      def decorate node
+        return unless @decorators
+        @decorators.each { |klass,list|
+          next unless node.is_a?(klass)
+          list.each { |moodule| node.extend(moodule) }
+        }
+      end
+
+      alias :to_xml :serialize
+      alias :clone :dup
+
+      # Get the hash of namespaces on the root Nokogiri::XML::Node
       def namespaces
-        root ? root.collect_namespaces : {}
+        root ? root.namespaces : {}
+      end
+
+      ##
+      # Create a Nokogiri::XML::DocumentFragment from +tags+
+      # Returns an empty fragment if +tags+ is nil.
+      def fragment tags = nil
+        DocumentFragment.new(self, tags, self.root)
+      end
+
+      undef_method :swap, :parent, :namespace, :default_namespace=
+      undef_method :add_namespace_definition, :attributes
+      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'
+        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
+          super(node_or_tags.first)
+        else
+          super
+        end
+      end
+      alias :<< :add_child
+
+      ##
+      # +JRuby+
+      # Wraps Java's org.w3c.dom.document and returns Nokogiri::XML::Document
+      def self.wrap document
+        raise "JRuby only method" unless Nokogiri.jruby?
+        return wrapJavaDocument(document)
+      end
+
+      ##
+      # +JRuby+
+      # Returns Java's org.w3c.dom.document of this Document.
+      def to_java
+        raise "JRuby only method" unless Nokogiri.jruby?
+        return toJavaDocument()
+      end
+
+      private
+      def self.empty_doc? string_or_io
+        string_or_io.nil? ||
+          (string_or_io.respond_to?(:empty?) && string_or_io.empty?) ||
+          (string_or_io.respond_to?(:eof?) && string_or_io.eof?)
+      end
+
+      def implied_xpath_contexts # :nodoc:
+        ["//"]
+      end
+
+      def inspect_attributes
+        [:name, :children]
       end
     end
   end

data/lib/nokogiri/xml/document_fragment.rb

@@ -0,0 +1,149 @@
+module Nokogiri
+  module XML
+    class DocumentFragment < Nokogiri::XML::Node
+      ##
+      #  Create a new DocumentFragment from +tags+.
+      #
+      #  If +ctx+ is present, it is used as a context node for the
+      #  subtree created, e.g., namespaces will be resolved relative
+      #  to +ctx+.
+      def initialize document, tags = nil, ctx = nil
+        return self unless tags
+
+        children = if ctx
+                     # Fix for issue#490
+                     if Nokogiri.jruby?
+                       # fix for issue #770
+                       ctx.parse("<root #{namespace_declarations(ctx)}>#{tags}</root>").children
+                     else
+                       ctx.parse(tags)
+                     end
+                   else
+                     XML::Document.parse("<root>#{tags}</root>") \
+                       .xpath("/root/node()")
+                   end
+        children.each { |child| child.parent = self }
+      end
+
+      ###
+      # return the name for DocumentFragment
+      def name
+        '#document-fragment'
+      end
+
+      ###
+      # Convert this DocumentFragment to a string
+      def to_s
+        children.to_s
+      end
+
+      ###
+      # Convert this DocumentFragment to html
+      # See Nokogiri::XML::NodeSet#to_html
+      def to_html *args
+        if Nokogiri.jruby?
+          options = args.first.is_a?(Hash) ? args.shift : {}
+          if !options[:save_with]
+            options[:save_with] = Node::SaveOptions::NO_DECLARATION | Node::SaveOptions::NO_EMPTY_TAGS | Node::SaveOptions::AS_HTML
+          end
+          args.insert(0, options)
+        end
+        children.to_html(*args)
+      end
+
+      ###
+      # Convert this DocumentFragment to xhtml
+      # See Nokogiri::XML::NodeSet#to_xhtml
+      def to_xhtml *args
+        if Nokogiri.jruby?
+          options = args.first.is_a?(Hash) ? args.shift : {}
+          if !options[:save_with]
+            options[:save_with] = Node::SaveOptions::NO_DECLARATION | Node::SaveOptions::NO_EMPTY_TAGS | Node::SaveOptions::AS_XHTML
+          end
+          args.insert(0, options)
+        end
+        children.to_xhtml(*args)
+      end
+
+      ###
+      # Convert this DocumentFragment to xml
+      # See Nokogiri::XML::NodeSet#to_xml
+      def to_xml *args
+        children.to_xml(*args)
+      end
+
+      ###
+      # 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' 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
+        ####
+        # Create a Nokogiri::XML::DocumentFragment from +tags+
+        def parse tags
+          self.new(XML::Document.new, tags)
+        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
+      def namespace_declarations ctx
+        ctx.namespace_scopes.map do |namespace|
+          prefix = namespace.prefix.nil? ? "" : ":#{namespace.prefix}"
+          %Q{xmlns#{prefix}="#{namespace.href}"}
+        end.join ' '
+      end
+
+      def coerce data
+        return super unless String === data
+
+        document.fragment(data).children
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/dtd.rb

@@ -1,6 +1,32 @@
 module Nokogiri
   module XML
-    class DTD < Node
+    class DTD < Nokogiri::XML::Node
+      undef_method :attribute_nodes
+      undef_method :values
+      undef_method :content
+      undef_method :namespace
+      undef_method :namespace_definitions
+      undef_method :line if method_defined?(:line)
+
+      def keys
+        attributes.keys
+      end
+
+      def each
+        attributes.each do |key, value|
+          yield([key, value])
+        end
+      end
+
+      def html_dtd?
+        name.casecmp('html').zero?
+      end
+
+      def html5_dtd?
+        html_dtd? &&
+          external_id.nil? &&
+          (system_id.nil? || system_id == 'about:legacy-compat')
+      end
     end
   end
 end

data/lib/nokogiri/xml/element_content.rb

@@ -0,0 +1,36 @@
+module Nokogiri
+  module XML
+    ###
+    # Represents the allowed content in an Element Declaration inside a DTD:
+    #
+    #   <?xml version="1.0"?><?TEST-STYLE PIDATA?>
+    #   <!DOCTYPE staff SYSTEM "staff.dtd" [
+    #      <!ELEMENT div1 (head, (p | list | note)*, div2*)>
+    #   ]>
+    #   </root>
+    #
+    # ElementContent represents the tree inside the <!ELEMENT> tag shown above
+    # that lists the possible content for the div1 tag.
+    class ElementContent
+      # Possible definitions of type
+      PCDATA  = 1
+      ELEMENT = 2
+      SEQ     = 3
+      OR      = 4
+
+      # Possible content occurrences
+      ONCE    = 1
+      OPT     = 2
+      MULT    = 3
+      PLUS    = 4
+
+      attr_reader :document
+
+      ###
+      # Get the children of this ElementContent node
+      def children
+        [c1, c2].compact
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/element_decl.rb

@@ -0,0 +1,13 @@
+module Nokogiri
+  module XML
+    class ElementDecl < Nokogiri::XML::Node
+      undef_method :namespace
+      undef_method :namespace_definitions
+      undef_method :line if method_defined?(:line)
+
+      def inspect
+        "#<#{self.class.name}:#{sprintf("0x%x", object_id)} #{to_s.inspect}>"
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/entity_decl.rb

@@ -0,0 +1,19 @@
+module Nokogiri
+  module XML
+    class EntityDecl < Nokogiri::XML::Node
+      undef_method :attribute_nodes
+      undef_method :attributes
+      undef_method :namespace
+      undef_method :namespace_definitions
+      undef_method :line if method_defined?(:line)
+
+      def self.new name, doc, *args
+        doc.create_entity(name, *args)
+      end
+
+      def inspect
+        "#<#{self.class.name}:#{sprintf("0x%x", object_id)} #{to_s.inspect}>"
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/namespace.rb

@@ -0,0 +1,13 @@
+module Nokogiri
+  module XML
+    class Namespace
+      include Nokogiri::XML::PP::Node
+      attr_reader :document
+
+      private
+      def inspect_attributes
+        [:prefix, :href]
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/node/save_options.rb

@@ -0,0 +1,61 @@
+module Nokogiri
+  module XML
+    class Node
+      ###
+      # Save options for serializing nodes
+      class SaveOptions
+        # Format serialized xml
+        FORMAT          = 1
+        # Do not include declarations
+        NO_DECLARATION  = 2
+        # Do not include empty tags
+        NO_EMPTY_TAGS   = 4
+        # Do not save XHTML
+        NO_XHTML        = 8
+        # Save as XHTML
+        AS_XHTML        = 16
+        # Save as XML
+        AS_XML          = 32
+        # Save as HTML
+        AS_HTML         = 64
+
+        if Nokogiri.jruby?
+          # Save builder created document
+          AS_BUILDER    = 128
+          # the default for XML documents
+          DEFAULT_XML  = AS_XML # https://github.com/sparklemotion/nokogiri/issues/#issue/415
+          # the default for HTML document
+          DEFAULT_HTML = NO_DECLARATION | NO_EMPTY_TAGS | AS_HTML
+        else
+          # the default for XML documents
+          DEFAULT_XML  = FORMAT | AS_XML
+          # the default for HTML document
+          DEFAULT_HTML = FORMAT | NO_DECLARATION | NO_EMPTY_TAGS | AS_HTML
+        end
+        # the default for XHTML document
+        DEFAULT_XHTML = FORMAT | NO_DECLARATION | NO_EMPTY_TAGS | AS_XHTML
+
+        # Integer representation of the SaveOptions
+        attr_reader :options
+
+        # Create a new SaveOptions object with +options+
+        def initialize options = 0; @options = options; end
+
+        constants.each do |constant|
+          class_eval %{
+            def #{constant.downcase}
+              @options |= #{constant}
+              self
+            end
+
+            def #{constant.downcase}?
+              #{constant} & @options == #{constant}
+            end
+          }
+        end
+
+        alias :to_i :options
+      end
+    end
+  end
+end