nokogiri 1.4.3.1-java → 1.5.0-java

data/lib/nokogiri/html.rb

@@ -4,6 +4,7 @@ require 'nokogiri/html/document_fragment'
 require 'nokogiri/html/sax/parser_context'
 require 'nokogiri/html/sax/parser'
 require 'nokogiri/html/element_description'
+require 'nokogiri/html/element_description_defaults'
 
 module Nokogiri
   class << self

data/lib/nokogiri/version.rb

@@ -1,38 +1,88 @@
 module Nokogiri
   # The version of Nokogiri you are using
-  VERSION = '1.4.3.1'
+  VERSION = '1.5.0'
 
-  # More complete version information about libxml
-  VERSION_INFO = {}
-  VERSION_INFO['warnings']              = []
-  VERSION_INFO['nokogiri']              = Nokogiri::VERSION
-  VERSION_INFO['ruby']                  = {}
-  VERSION_INFO['ruby']['version']       = ::RUBY_VERSION
-  VERSION_INFO['ruby']['platform']      = ::RUBY_PLATFORM
-  VERSION_INFO['ruby']['engine']        = defined?(RUBY_ENGINE) ? RUBY_ENGINE : 'mri'
-  VERSION_INFO['ruby']['jruby']         = ::JRUBY_VERSION if RUBY_PLATFORM == "java"
-  if defined?(LIBXML_VERSION)
-    VERSION_INFO['libxml']              = {}
-    VERSION_INFO['libxml']['binding']   = 'extension'
-    VERSION_INFO['libxml']['compiled']  = LIBXML_VERSION
-    VERSION_INFO['libxml']['loaded']    = LIBXML_PARSER_VERSION.scan(/^(.*)(..)(..)$/).first.collect{|j|j.to_i}.join(".")
-
-    if VERSION_INFO['libxml']['compiled'] != VERSION_INFO['libxml']['loaded']
-      warning = "Nokogiri was built against LibXML version #{VERSION_INFO['libxml']['compiled']}, but has dynamically loaded #{VERSION_INFO['libxml']['loaded']}"
-      VERSION_INFO['warnings'] << warning
+  class VersionInfo # :nodoc:
+    def jruby?
+      ::JRUBY_VERSION if RUBY_PLATFORM == "java"
+    end
+
+    def engine
+      defined?(RUBY_ENGINE) ? RUBY_ENGINE : 'mri'
+    end
+
+    def loaded_parser_version
+      LIBXML_PARSER_VERSION.scan(/^(.*)(..)(..)$/).first.collect{ |j|
+        j.to_i
+      }.join(".")
+    end
+
+    def compiled_parser_version
+      LIBXML_VERSION
+    end
+
+    def libxml2?
+      defined?(LIBXML_VERSION)
+    end
+
+    def warnings
+      return [] unless libxml2?
+
+      if compiled_parser_version != loaded_parser_version
+        ["Nokogiri was built against LibXML version #{compiled_parser_version}, but has dynamically loaded #{loaded_parser_version}"]
+      else
+        []
+      end
+    end
+
+    def to_hash
+      hash_info = {}
+      hash_info['warnings']              = []
+      hash_info['nokogiri']              = Nokogiri::VERSION
+      hash_info['ruby']                  = {}
+      hash_info['ruby']['version']       = ::RUBY_VERSION
+      hash_info['ruby']['platform']      = ::RUBY_PLATFORM
+      hash_info['ruby']['description']   = ::RUBY_DESCRIPTION
+      hash_info['ruby']['engine']        = engine
+      hash_info['ruby']['jruby']         = jruby? if jruby?
+
+      if libxml2?
+        hash_info['libxml']              = {}
+        hash_info['libxml']['binding']   = 'extension'
+        hash_info['libxml']['compiled']  = compiled_parser_version
+        hash_info['libxml']['loaded']    = loaded_parser_version
+        hash_info['warnings']            = warnings
+      end
+
+      hash_info
+    end
+
+    def to_markdown
+      begin
+        require 'psych'
+      rescue LoadError
+      end
+      require 'yaml'
+      "# Nokogiri (#{Nokogiri::VERSION})\n" +
+      YAML.dump(to_hash).each_line.map { |line| "    #{line}" }.join
+    end
+
+    # FIXME: maybe switch to singleton?
+    @@instance = new
+    @@instance.warnings.each do |warning|
       warn "WARNING: #{warning}"
     end
+    def self.instance; @@instance; end
   end
 
-  def self.uses_libxml? # :nodoc:
-    !Nokogiri::VERSION_INFO['libxml'].nil?
-  end
+  # More complete version information about libxml
+  VERSION_INFO = VersionInfo.instance.to_hash
 
-  def self.ffi? # :nodoc:
-    uses_libxml? && Nokogiri::VERSION_INFO['libxml']['binding'] == 'ffi'
+  def self.uses_libxml? # :nodoc:
+    VersionInfo.instance.libxml2?
   end
 
-  def self.is_2_6_16? # :nodoc:
-    Nokogiri::VERSION_INFO['libxml']['loaded'] <= '2.6.16'
+  def self.jruby? # :nodoc:
+    VersionInfo.instance.jruby?
   end
 end

data/lib/nokogiri/xml/attribute_decl.rb

@@ -8,7 +8,7 @@ module Nokogiri
       undef_method :content
       undef_method :namespace
       undef_method :namespace_definitions
-      undef_method :line
+      undef_method :line if method_defined?(:line)
 
       def inspect
         "#<#{self.class.name}:#{sprintf("0x%x", object_id)} #{to_s.inspect}>"

data/lib/nokogiri/xml/builder.rb

@@ -327,6 +327,13 @@ module Nokogiri
       ###
       # Convert this Builder object to XML
       def to_xml(*args)
+        if Nokogiri.jruby?
+          options = args.first.is_a?(Hash) ? args.shift : {}
+          if !options[:save_with]
+            options[:save_with] = Node::SaveOptions::AS_BUILDER
+          end
+          args.insert(0, options)
+        end
         @doc.to_xml(*args)
       end
 

data/lib/nokogiri/xml/document.rb

@@ -9,7 +9,7 @@ module Nokogiri
     # Nokogiri::XML::Node#xpath
     class Document < Nokogiri::XML::Node
       ##
-      # Parse an XML file.  +thing+ may be a String, or any object that
+      # 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+ is resource where this document is located.  +encoding+ is the
       # encoding that should be used when processing the document. +options+
@@ -113,6 +113,13 @@ module Nokogiri
       #   </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 this is a very expensive operation in current implementation, as it
+      # traverses the entire graph, and also has to bring each node accross the
+      # libxml bridge into a ruby object.
       def collect_namespaces
         ns = {}
         traverse { |j| ns.merge!(j.namespaces) }
@@ -134,7 +141,21 @@ module Nokogiri
       end
 
       ##
-      # Explore a document with shortcut methods.
+      # 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
@@ -184,7 +205,27 @@ module Nokogiri
       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 implied_xpath_context
+        "/"
+      end
+
       def inspect_attributes
         [:name, :children]
       end

data/lib/nokogiri/xml/document_fragment.rb

@@ -11,9 +11,9 @@ module Nokogiri
         return self unless tags
 
         children = if ctx
-                     ctx.parse(tags.strip)
+                     ctx.parse(tags)
                    else
-                     XML::Document.parse("<root>#{tags.strip}</root>") \
+                     XML::Document.parse("<root>#{tags}</root>") \
                        .xpath("/root/node()")
                    end
         children.each { |child| child.parent = self }
@@ -35,6 +35,13 @@ module Nokogiri
       # 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
 
@@ -42,6 +49,13 @@ module Nokogiri
       # 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
 

data/lib/nokogiri/xml/dtd.rb

@@ -2,10 +2,21 @@ module Nokogiri
   module XML
     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
+      undef_method :line if method_defined?(:line)
+
+      def keys
+        attributes.keys
+      end
+
+      def each &block
+        attributes.each { |key, value|
+          block.call([key, value])
+        }
+      end
     end
   end
 end

data/lib/nokogiri/xml/element_decl.rb

@@ -3,7 +3,7 @@ module Nokogiri
     class ElementDecl < Nokogiri::XML::Node
       undef_method :namespace
       undef_method :namespace_definitions
-      undef_method :line
+      undef_method :line if method_defined?(:line)
 
       def inspect
         "#<#{self.class.name}:#{sprintf("0x%x", object_id)} #{to_s.inspect}>"

data/lib/nokogiri/xml/entity_decl.rb

@@ -5,7 +5,7 @@ module Nokogiri
       undef_method :attributes
       undef_method :namespace
       undef_method :namespace_definitions
-      undef_method :line
+      undef_method :line if method_defined?(:line)
 
       def self.new name, doc, *args
         doc.create_entity(name, *args)

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

@@ -6,7 +6,7 @@ module Nokogiri
       class SaveOptions
         # Format serialized xml
         FORMAT          = 1
-        # Do not include delcarations
+        # Do not include declarations
         NO_DECLARATION  = 2
         # Do not include empty tags
         NO_EMPTY_TAGS   = 4
@@ -19,11 +19,28 @@ module Nokogiri
         # 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/tenderlove/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}
@@ -36,6 +53,8 @@ module Nokogiri
             end
           }
         end
+
+        alias :to_i :options
       end
     end
   end