nokogiri 1.4.7 → 1.5.0.beta.1

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

@@ -19,19 +19,11 @@ module Nokogiri
         # Save as HTML
         AS_HTML         = 64
 
-        # 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
-        # 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}
@@ -44,8 +36,6 @@ module Nokogiri
             end
           }
         end
-
-        alias :to_i :options
       end
     end
   end

data/lib/nokogiri/xml/node_set.rb

@@ -42,7 +42,7 @@ module Nokogiri
       end
 
       ###
-      # Returns the index of the first node in self that is == to +node+. Returns nil if no match is found.
+      # Returns the index of the first node in self that is == to +node+. Returns nil if no match is found. 
       def index(node)
         each_with_index { |member, j| return j if member == node }
         nil
@@ -104,7 +104,7 @@ module Nokogiri
 
         each do |node|
           doc = node.document
-          search_ns = ns || (doc.root ? doc.root.namespaces : {})
+          search_ns = ns || doc.root ? doc.root.namespaces : {}
 
           xpaths = paths.map { |rule|
             [
@@ -138,13 +138,6 @@ module Nokogiri
         sub_set
       end
 
-      ###
-      # Search this NodeSet's nodes' immediate children using CSS selector +selector+
-      def > selector
-        ns = document.root.namespaces
-        xpath CSS.xpath_for(selector, :prefix => "./", :ns => ns).first
-      end
-
       ###
       # If path is a string, search this document for +path+ returning the
       # first Node.  Otherwise, index in to the array with +path+.
@@ -154,24 +147,6 @@ module Nokogiri
       end
       alias :% :at
 
-      ##
-      # Search this NodeSet for the first occurrence of XPath +paths+.
-      # Equivalent to <tt>xpath(paths).first</tt>
-      # See NodeSet#xpath for more information.
-      #
-      def at_xpath *paths
-        xpath(*paths).first
-      end
-
-      ##
-      # Search this NodeSet for the first occurrence of CSS +rules+.
-      # Equivalent to <tt>css(rules).first</tt>
-      # See NodeSet#css for more information.
-      #
-      def at_css *rules
-        css(*rules).first
-      end
-
       ###
       # Filter this list for nodes that match +expr+
       def filter expr
@@ -257,7 +232,7 @@ module Nokogiri
       # Wrap this NodeSet with +html+ or the results of the builder in +blk+
       def wrap(html, &blk)
         each do |j|
-          new_parent = document.parse(html).first
+          new_parent = document.root.parse(html).first
           j.add_next_sibling(new_parent)
           new_parent.add_child(j)
         end

data/lib/nokogiri/xml/parse_options.rb

@@ -37,14 +37,6 @@ module Nokogiri
       NOCDATA     = 1 << 14
       # do not generate XINCLUDE START/END nodes
       NOXINCNODE  = 1 << 15
-      # compact small text nodes; no modification of the tree allowed afterwards (will possibly crash if you try to modify the tree)
-      COMPACT     = 1 << 16
-      # parse using XML-1.0 before update 5
-      OLD10       = 1 << 17
-      # do not fixup XINCLUDE xml:base uris
-      NOBASEFIX   = 1 << 18
-      # relax any hardcoded limit from the parser
-      HUGE        = 1 << 19
 
       # the default options used for parsing XML documents
       DEFAULT_XML  = RECOVER

data/lib/nokogiri/xml/reader.rb

@@ -30,42 +30,6 @@ module Nokogiri
     class Reader
       include Enumerable
 
-      TYPE_NONE = 0
-      # Element node type
-      TYPE_ELEMENT = 1
-      # Attribute node type
-      TYPE_ATTRIBUTE = 2
-      # Text node type
-      TYPE_TEXT = 3
-      # CDATA node type
-      TYPE_CDATA = 4
-      # Entity Reference node type
-      TYPE_ENTITY_REFERENCE = 5
-      # Entity node type
-      TYPE_ENTITY = 6
-      # PI node type
-      TYPE_PROCESSING_INSTRUCTION = 7
-      # Comment node type
-      TYPE_COMMENT = 8
-      # Document node type
-      TYPE_DOCUMENT = 9
-      # Document Type node type
-      TYPE_DOCUMENT_TYPE = 10
-      # Document Fragment node type
-      TYPE_DOCUMENT_FRAGMENT = 11
-      # Notation node type
-      TYPE_NOTATION = 12
-      # Whitespace node type
-      TYPE_WHITESPACE = 13
-      # Significant Whitespace node type
-      TYPE_SIGNIFICANT_WHITESPACE = 14
-      # Element end node type
-      TYPE_END_ELEMENT = 15
-      # Entity end node type
-      TYPE_END_ENTITY = 16
-      # XML Declaration node type
-      TYPE_XML_DECLARATION = 17
-
       # A list of errors encountered while parsing
       attr_accessor :errors
 
@@ -75,8 +39,6 @@ module Nokogiri
       # The XML source
       attr_reader :source
 
-      alias :self_closing? :empty_element?
-
       def initialize source, url = nil, encoding = nil # :nodoc:
         @source   = source
         @errors   = []
@@ -87,9 +49,9 @@ module Nokogiri
       ###
       # Get a list of attributes for the current node.
       def attributes
-        Hash[*attribute_nodes.map { |node|
+        Hash[*(attribute_nodes.map { |node|
           [node.name, node.to_s]
-        }.flatten].merge(namespaces || {})
+        }.flatten)].merge(namespaces || {})
       end
 
       ###
@@ -101,10 +63,10 @@ module Nokogiri
       end
 
       ###
-      # Move the cursor through the document yielding the cursor to the block
-      def each
-        while cursor = self.read
-          yield cursor
+      # Move the cursor through the document yielding each node to the block
+      def each(&block)
+        while node = self.read
+          block.call(node)
         end
       end
     end

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

@@ -8,7 +8,7 @@ module Nokogiri
     # The basic way a SAX style parser works is by creating a parser,
     # telling the parser about the events we're interested in, then giving
     # the parser some XML to process.  The parser will notify you when
-    # it encounters events you said you would like to know about.
+    # it encounters events your said you would like to know about.
     #
     # To register for events, you simply subclass Nokogiri::XML::SAX::Document,
     # and implement the methods for which you would like notification.
@@ -36,7 +36,7 @@ module Nokogiri
     #   parser.parse(File.read(ARGV[0], 'rb'))
     #
     # Now my document handler will be called when each node starts, and when
-    # the document ends.  To see what kinds of events are available, take
+    # then document ends.  To see what kinds of events are available, take
     # a look at Nokogiri::XML::SAX::Document.
     #
     # Two SAX parsers for XML are available, a parser that reads from a string
@@ -85,9 +85,7 @@ module Nokogiri
 
         ###
         # Called at the beginning of an element
-        # * +name+ is the name of the tag
-        # * +attrs+ are an assoc list of namespaces and attributes, e.g.:
-        #     [ ["xmlns:foo", "http://sample.net"], ["size", "large"] ]
+        # +name+ is the name of the tag with +attrs+ as attributes
         def start_element name, attrs = []
         end
 
@@ -112,7 +110,7 @@ module Nokogiri
             [['xmlns', ns_prefix].compact.join(':'), ns_uri]
           } + attrs.map { |attr|
             [[attr.prefix, attr.localname].compact.join(':'), attr.value]
-          }
+          }.flatten
           start_element name, attributes
         end
 
@@ -128,9 +126,7 @@ module Nokogiri
         end
 
         ###
-        # Characters read between a tag.  This method might be called multiple
-        # times given one contiguous string of characters.
-        #
+        # Characters read between a tag
         # +string+ contains the character data
         def characters string
         end

data/lib/nokogiri/xml/schema.rb

@@ -43,13 +43,7 @@ module Nokogiri
       # Nokogiri::XML::SyntaxError objects found while validating the
       # +thing+ is returned.
       def validate thing
-        if thing.is_a?(Nokogiri::XML::Document) 
-          validate_document(thing) 
-        elsif File.file?(thing)
-          validate_file(thing)
-        else
-          raise ArgumentError, "Must provide Nokogiri::Xml::Document or the name of an existing file"
-        end
+        thing.is_a?(Nokogiri::XML::Document) ? validate_document(thing) : validate_file(thing)
       end
 
       ###

data/lib/nokogiri/xslt.rb

@@ -9,8 +9,8 @@ module Nokogiri
     #
     #   xslt = Nokogiri::XSLT(File.read(ARGV[0]))
     #
-    def XSLT stylesheet, modules = {}
-      XSLT.parse(stylesheet, modules)
+    def XSLT stylesheet
+      XSLT.parse(stylesheet)
     end
   end
 
@@ -20,15 +20,11 @@ module Nokogiri
   module XSLT
     class << self
       ###
-      # Parse the stylesheet in +string+, register any +modules+
-      def parse string, modules = {}
-        modules.each do |url, klass|
-          XSLT.register url, klass
-	end
-
+      # Parse the stylesheet in +string+
+      def parse string
         Stylesheet.parse_stylesheet_doc(XML.parse(string))
       end
-
+      
       ###
       # Quote parameters in +params+ for stylesheet safety
       def quote_params params

data/tasks/cross_compile.rb

@@ -6,7 +6,6 @@ ZLIB    = 'zlib-1.2.5'
 ICONV   = 'libiconv-1.13.1'
 LIBXML  = 'libxml2-2.7.7'
 LIBXSLT = 'libxslt-1.1.26'
-RAKE_COMPILER_PKGCONFIG = File.expand_path(File.join(Dir.pwd, "tmp/cross/lib/pkgconfig/"))
 
 ### Build zlib ###
 file "tmp/cross/download/#{ZLIB}" do |t|
@@ -27,6 +26,7 @@ file "tmp/cross/download/#{ZLIB}" do |t|
       f.puts "LIBRARY_PATH = #{CROSS_DIR}/lib"
       f.puts "INCLUDE_PATH = #{CROSS_DIR}/include"
 
+      # FIXME: need to make the cross compiler dynamic
       f.puts mk.sub(/^PREFIX\s*=\s*$/, "PREFIX = #{HOST}-") #.
         #sub(/^SHARED_MODE=0$/, 'SHARED_MODE=1').
         #sub(/^IMPLIB\s*=.*$/, 'IMPLIB=libz.dll.a')
@@ -55,6 +55,7 @@ file "tmp/cross/download/#{ICONV}" do |t|
   end
 
   Dir.chdir t.name do
+    # FIXME: need to make the host dynamic
     sh "./configure --disable-shared --enable-static --host=#{HOST} --target=#{TARGET} --prefix=#{CROSS_DIR} CPPFLAGS='-mno-cygwin -Wall' CFLAGS='-mno-cygwin -O2 -g' CXXFLAGS='-mno-cygwin -O2 -g' LDFLAGS=-mno-cygwin"
   end
 end
@@ -80,6 +81,7 @@ file "tmp/cross/download/#{LIBXML}" do |t|
   end
 
   Dir.chdir t.name do
+    # FIXME: need to make the host dynamic
     sh "CFLAGS='-DIN_LIBXML' ./configure --host=#{HOST} --target=#{TARGET} --enable-static --disable-shared --prefix=#{CROSS_DIR} --with-zlib=#{CROSS_DIR} --with-iconv=#{CROSS_DIR} --without-python --without-readline"
   end
 end
@@ -105,7 +107,8 @@ file "tmp/cross/download/#{LIBXSLT}" do |t|
   end
 
   Dir.chdir t.name do
-    sh "CFLAGS='-DIN_LIBXML' ./configure --host=#{HOST} --target=#{TARGET} --enable-static --disable-shared --prefix=#{CROSS_DIR} --with-libxml-prefix=#{CROSS_DIR} --without-python --without-crypto"
+    # FIXME: need to make the host dynamic
+    sh "CFLAGS='-DIN_LIBXML' ./configure --host=#{HOST} --target=#{TARGET} --enable-static --disable-shared --prefix=#{CROSS_DIR} --with-libxml-prefix=#{CROSS_DIR} --without-python"
   end
 end
 
@@ -117,7 +120,7 @@ file 'tmp/cross/bin/xslt-config' => "tmp/cross/download/#{LIBXSLT}" do |t|
 end
 ### End build libxslt ###
 
-file 'lib/nokogiri/nokogiri.rb' => 'cross:check' do
+file 'lib/nokogiri/nokogiri.rb' => 'cross:libxslt' do
   File.open("lib/#{HOE.name}/#{HOE.name}.rb", 'wb') do |f|
     f.write <<-eoruby
 require "#{HOE.name}/\#{RUBY_VERSION.sub(/\\.\\d+$/, '')}/#{HOE.name}"
@@ -131,43 +134,25 @@ namespace :cross do
   task :libxml2 => ['cross:zlib', 'cross:iconv', 'tmp/cross/bin/xml2-config']
   task :libxslt => ['cross:libxml2', 'tmp/cross/bin/xslt-config']
 
-  task :check => ["cross:libxslt"] do
-    unless File.directory?(RAKE_COMPILER_PKGCONFIG)
-      raise RuntimeError.new("looks like rake-compiler changed where pkgconfig info is kept. (#{RAKE_COMPILER_PKGCONFIG})")
+  task :copy_dlls do
+    Dir['tmp/cross/bin/*.dll'].each do |file|
+      cp file, "ext/nokogiri"
     end
   end
 
-  task :file_list do
+  task :file_list => 'cross:copy_dlls' do
     HOE.spec.extensions = []
     HOE.spec.files += Dir["lib/#{HOE.name}/#{HOE.name}.rb"]
     HOE.spec.files += Dir["lib/#{HOE.name}/1.{8,9}/#{HOE.name}.so"]
+    HOE.spec.files += Dir["ext/nokogiri/*.dll"]
   end
 end
 
 CLOBBER.include("lib/nokogiri/nokogiri.{so,dylib,rb,bundle}")
 CLOBBER.include("lib/nokogiri/1.{8,9}")
+CLOBBER.include("ext/nokogiri/*.dll")
 
 if Rake::Task.task_defined?(:cross)
   Rake::Task[:cross].prerequisites << "lib/nokogiri/nokogiri.rb"
   Rake::Task[:cross].prerequisites << "cross:file_list"
 end
-
-desc "build a windows gem without all the ceremony."
-task "gem:windows" do
-  rake_compiler_config = YAML.load_file("#{ENV['HOME']}/.rake-compiler/config.yml")
-
-  # check that rake-compiler config contains the right patchlevels of 1.8.6 and 1.9.1. see #279.
-  ["1.8.6-p383", "1.9.1-p243"].each do |version|
-    majmin, patchlevel = version.split("-")
-    rbconfig = "rbconfig-#{majmin}"
-    unless rake_compiler_config.key?(rbconfig) && rake_compiler_config[rbconfig] =~ /-#{patchlevel}/
-      raise "rake-compiler '#{rbconfig}' not #{patchlevel}. try running 'env --unset=HOST rake-compiler cross-ruby VERSION=#{version}'"
-    end
-  end
-
-  # verify that --export-all is in the 1.9.1 rbconfig. see #279,#374,#375.
-  rbconfig_191 = rake_compiler_config["rbconfig-1.9.1"]
-  raise "rbconfig #{rbconfig_191} needs --export-all in its DLDFLAGS value" if File.read(rbconfig_191).grep(/CONFIG\["DLDFLAGS"\].*--export-all/).empty?
-
-  system("env PKG_CONFIG_PATH=#{RAKE_COMPILER_PKGCONFIG} RUBY_CC_VERSION=1.8.6:1.9.1 rake cross native gem") || raise("build failed!")
-end

data/test/css/test_parser.rb

@@ -16,10 +16,6 @@ module Nokogiri
         assert_raises(CSS::SyntaxError) { @parser.parse("a[x=]") }
       end
 
-      def test_function_and_pseudo
-        assert_xpath '//child::text()[position() = 99]', @parser.parse('text():nth-of-type(99)')
-      end
-
       def test_find_by_type
         ast = @parser.parse("a:nth-child(2)").first
         matches = ast.find_by_type(
@@ -142,31 +138,30 @@ module Nokogiri
 
       def test_nonstandard_nth_selectors
         ## These are non standard CSS
-        assert_xpath '//a[position() = 1]',             @parser.parse('a:first()')
-        assert_xpath '//a[position() = 1]',             @parser.parse('a:first') # no parens
-        assert_xpath '//a[position() = 99]',            @parser.parse('a:eq(99)')
-        assert_xpath '//a[position() = 99]',            @parser.parse('a:nth(99)')
-        assert_xpath '//a[position() = last()]',        @parser.parse('a:last()')
-        assert_xpath '//a[position() = last()]',        @parser.parse('a:last') # no parens
-        assert_xpath '//a[node()]',                     @parser.parse('a:parent')
+        assert_xpath '//a[position() = 99]', @parser.parse('a:eq(99)')
+        assert_xpath '//a[position() = 1]', @parser.parse('a:first') # no parens
+        assert_xpath '//a[position() = last()]', @parser.parse('a:last') # no parens
+        assert_xpath '//a[position() = 99]', @parser.parse('a:nth(99)')
+        assert_xpath '//a[position() = 1]', @parser.parse('a:first()')
+        assert_xpath '//a[position() = last()]', @parser.parse('a:last()')
+        assert_xpath '//a[node()]', @parser.parse('a:parent')
       end
 
       def test_standard_nth_selectors
-        assert_xpath '//a[position() = 1]',             @parser.parse('a:first-of-type()')
-        assert_xpath '//a[position() = 1]',             @parser.parse('a:first-of-type') # no parens
-        assert_xpath '//a[position() = 99]',            @parser.parse('a:nth-of-type(99)')
-        assert_xpath '//a[position() = last()]',        @parser.parse('a:last-of-type()')
-        assert_xpath '//a[position() = last()]',        @parser.parse('a:last-of-type') # no parens
-        assert_xpath '//a[position() = last()]',        @parser.parse('a:nth-last-of-type(1)')
-        assert_xpath '//a[position() = last() - 98]',   @parser.parse('a:nth-last-of-type(99)')
+        assert_xpath '//a[position() = 99]', @parser.parse('a:nth-of-type(99)')
+        assert_xpath '//a[position() = 1]', @parser.parse('a:first-of-type()')
+        assert_xpath '//a[position() = last()]', @parser.parse('a:last-of-type()')
+        assert_xpath '//a[position() = 1]', @parser.parse('a:first-of-type') # no parens
+        assert_xpath '//a[position() = last()]', @parser.parse('a:last-of-type') # no parens
+        assert_xpath '//a[position() = last() - 99]', @parser.parse('a:nth-last-of-type(99)')
+        assert_xpath '//a[position() = last() - 99]', @parser.parse('a:nth-last-of-type(99)')
       end
 
       def test_nth_child_selectors
-        assert_xpath '//*[position() = 1 and self::a]',           @parser.parse('a:first-child')
-        assert_xpath '//*[position() = 99 and self::a]',          @parser.parse('a:nth-child(99)')
-        assert_xpath '//*[position() = last() and self::a]',      @parser.parse('a:last-child')
-        assert_xpath '//*[position() = last() and self::a]',      @parser.parse('a:nth-last-child(1)')
-        assert_xpath '//*[position() = last() - 98 and self::a]', @parser.parse('a:nth-last-child(99)')
+        assert_xpath '//*[position() = 1 and self::a]', @parser.parse('a:first-child')
+        assert_xpath '//*[position() = last() and self::a]', @parser.parse('a:last-child')
+        assert_xpath '//*[position() = 99 and self::a]', @parser.parse('a:nth-child(99)')
+        assert_xpath '//*[position() = last() - 99 and self::a]', @parser.parse('a:nth-last-child(99)')
       end
 
       def test_miscellaneous_selectors
@@ -186,32 +181,16 @@ module Nokogiri
         assert_xpath '//a[(position() <= 3) and (((position()-3) mod 1) = 0)]', @parser.parse('a:nth-of-type(-n+3)')
         assert_xpath '//a[(position() >= 3) and (((position()-3) mod 1) = 0)]', @parser.parse('a:nth-of-type(1n+3)')
         assert_xpath '//a[(position() >= 3) and (((position()-3) mod 1) = 0)]', @parser.parse('a:nth-of-type(n+3)')
-
-        assert_xpath '//a[((last()-position()+1) mod 2) = 0]', @parser.parse('a:nth-last-of-type(2n)')
-        assert_xpath '//a[((last()-position()+1) >= 1) and ((((last()-position()+1)-1) mod 2) = 0)]', @parser.parse('a:nth-last-of-type(2n+1)')
-        assert_xpath '//a[((last()-position()+1) mod 2) = 0]', @parser.parse('a:nth-last-of-type(even)')
-        assert_xpath '//a[((last()-position()+1) >= 1) and ((((last()-position()+1)-1) mod 2) = 0)]', @parser.parse('a:nth-last-of-type(odd)')
-        assert_xpath '//a[((last()-position()+1) >= 3) and ((((last()-position()+1)-3) mod 4) = 0)]', @parser.parse('a:nth-last-of-type(4n+3)')
-        assert_xpath '//a[((last()-position()+1) <= 3) and ((((last()-position()+1)-3) mod 1) = 0)]', @parser.parse('a:nth-last-of-type(-1n+3)')
-        assert_xpath '//a[((last()-position()+1) <= 3) and ((((last()-position()+1)-3) mod 1) = 0)]', @parser.parse('a:nth-last-of-type(-n+3)')
-        assert_xpath '//a[((last()-position()+1) >= 3) and ((((last()-position()+1)-3) mod 1) = 0)]', @parser.parse('a:nth-last-of-type(1n+3)')
-        assert_xpath '//a[((last()-position()+1) >= 3) and ((((last()-position()+1)-3) mod 1) = 0)]', @parser.parse('a:nth-last-of-type(n+3)')
       end
 
       def test_preceding_selector
-        assert_xpath  "//E/following-sibling::F",
+        assert_xpath  "//F[preceding-sibling::E]",
                       @parser.parse("E ~ F")
-
-        assert_xpath  "//E/following-sibling::F//G",
-                      @parser.parse("E ~ F G")
       end
 
       def test_direct_preceding_selector
         assert_xpath  "//E/following-sibling::*[1]/self::F",
                       @parser.parse("E + F")
-
-        assert_xpath  "//E/following-sibling::*[1]/self::F//G",
-                      @parser.parse("E + F G")
       end
 
       def test_attribute