nokogiri 1.5.0.beta.2 → 1.5.0.beta.3

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
@@ -257,7 +257,7 @@ 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.root.parse(html).first
+          new_parent = document.parse(html).first
           j.add_next_sibling(new_parent)
           new_parent.add_child(j)
         end

data/lib/nokogiri/xml/reader.rb

@@ -30,6 +30,42 @@ 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
 

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

@@ -85,7 +85,9 @@ module Nokogiri
 
         ###
         # Called at the beginning of an element
-        # +name+ is the name of the tag with +attrs+ as attributes
+        # * +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"] ]
         def start_element name, attrs = []
         end
 
@@ -110,7 +112,7 @@ 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
 

data/lib/nokogiri/xslt.rb

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

data/lib/nokogiri/xslt/stylesheet.rb

@@ -6,7 +6,7 @@ module Nokogiri
     # an XML::Document with a Stylesheet:
     #
     #   doc   = Nokogiri::XML(File.read('some_file.xml'))
-    #   xslt  = Nokogir::XSLT(File.read('some_transformer.xslt'))
+    #   xslt  = Nokogiri::XSLT(File.read('some_transformer.xslt'))
     #
     #   puts xslt.transform(doc)
     #

data/tasks/cross_compile.rb

@@ -6,6 +6,7 @@ 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|
@@ -108,7 +109,7 @@ file "tmp/cross/download/#{LIBXSLT}" do |t|
 
   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-libxml-prefix=#{CROSS_DIR} --without-python"
+    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"
   end
 end
 
@@ -120,7 +121,7 @@ file 'tmp/cross/bin/xslt-config' => "tmp/cross/download/#{LIBXSLT}" do |t|
 end
 ### End build libxslt ###
 
-file 'lib/nokogiri/nokogiri.rb' => 'cross:libxslt' do
+file 'lib/nokogiri/nokogiri.rb' => 'cross:check' do
   File.open("lib/#{HOE.name}/#{HOE.name}.rb", 'wb') do |f|
     f.write <<-eoruby
 require "#{HOE.name}/\#{RUBY_VERSION.sub(/\\.\\d+$/, '')}/#{HOE.name}"
@@ -134,25 +135,43 @@ namespace :cross do
   task :libxml2 => ['cross:zlib', 'cross:iconv', 'tmp/cross/bin/xml2-config']
   task :libxslt => ['cross:libxml2', 'tmp/cross/bin/xslt-config']
 
-  task :copy_dlls do
-    Dir['tmp/cross/bin/*.dll'].each do |file|
-      cp file, "ext/nokogiri"
+  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})")
     end
   end
 
-  task :file_list => 'cross:copy_dlls' do
+  task :file_list 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 '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

@@ -142,30 +142,31 @@ module Nokogiri
 
       def test_nonstandard_nth_selectors
         ## These are non standard CSS
-        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')
+        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')
       end
 
       def test_standard_nth_selectors
-        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)')
+        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)')
       end
 
       def test_nth_child_selectors
-        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)')
+        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)')
       end
 
       def test_miscellaneous_selectors
@@ -185,6 +186,16 @@ 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

data/test/decorators/test_slop.rb

@@ -0,0 +1,16 @@
+require "helper"
+
+module Nokogiri
+  class TestSlop < Nokogiri::TestCase
+    def test_description_tag
+      doc = Nokogiri.Slop(<<-eoxml)
+        <item>
+          <title>foo</title>
+          <description>this is the foo thing</description>
+        </item>
+      eoxml
+      assert doc.item.title
+      assert doc.item._description, 'should have description'
+    end
+  end
+end

data/test/html/test_document_fragment.rb

@@ -74,6 +74,16 @@ module Nokogiri
         assert Nokogiri::HTML::DocumentFragment.new(@html)
       end
 
+      def test_body_fragment_should_contain_body
+        fragment = Nokogiri::HTML::DocumentFragment.parse("  <body><div>foo</div></body>")
+        assert_match(/^<body>/, fragment.to_s)
+      end
+
+      def test_nonbody_fragment_should_not_contain_body
+        fragment = Nokogiri::HTML::DocumentFragment.parse("<div>foo</div>")
+        assert_match(/^<div>/, fragment.to_s)
+      end
+
       def test_fragment_should_have_document
         fragment = Nokogiri::HTML::DocumentFragment.new(@html)
         assert_equal @html, fragment.document
@@ -140,13 +150,13 @@ module Nokogiri
       def test_html_fragment_with_leading_whitespace
         doc = "     <div>b</div>  "
         fragment = Nokogiri::HTML::Document.new.fragment(doc)
-        assert_equal "<div>b</div>", fragment.to_s
+        assert_match %r%     <div>b</div> *%, fragment.to_s
       end
 
       def test_html_fragment_with_leading_whitespace_and_newline
         doc = "     \n<div>b</div>  "
         fragment = Nokogiri::HTML::Document.new.fragment(doc)
-        assert_equal "<div>b</div>", fragment.to_s
+        assert_match %r%     \n<div>b</div> *%, fragment.to_s
       end
 
       def test_html_fragment_with_leading_text_and_newline
@@ -156,7 +166,7 @@ module Nokogiri
 
       def test_html_fragment_with_leading_whitespace_and_text_and_newline
         fragment = HTML::Document.new.fragment("  First line\nSecond line<br>Broken line")
-        assert_equal "First line\nSecond line<br>Broken line", fragment.to_s
+        assert_equal "  First line\nSecond line<br>Broken line", fragment.to_s
       end
 
       def test_html_fragment_with_leading_entity
@@ -207,6 +217,39 @@ module Nokogiri
         assert_equal("<p>hello<!-- your ad here --></p>",
           fragment.to_s)
       end
+
+      def test_malformed_fragment_is_corrected
+        fragment = HTML::DocumentFragment.parse("<div </div>")
+        assert_equal "<div></div>", fragment.to_s
+      end
+
+      def test_unclosed_script_tag
+        # see GH#315
+        fragment = HTML::DocumentFragment.parse("foo <script>bar")
+        assert_equal "foo <script>bar</script>", fragment.to_html
+      end
+
+      def test_error_propagation_on_fragment_parse
+        frag = Nokogiri::HTML::DocumentFragment.parse "<hello>oh, hello there.</hello>"
+        assert frag.errors.any?{|err| err.to_s =~ /Tag hello invalid/}, "errors should be copied to the fragment"
+      end
+
+      def test_error_propagation_on_fragment_parse_in_node_context
+        doc = Nokogiri::HTML::Document.parse "<html><body><div></div></body></html>"
+        context_node = doc.at_css "div"
+        frag = Nokogiri::HTML::DocumentFragment.new doc, "<hello>oh, hello there.</hello>", context_node
+        assert frag.errors.any?{|err| err.to_s =~ /Tag hello invalid/}, "errors should be on the context node's document"
+      end
+
+      def test_error_propagation_on_fragment_parse_in_node_context_should_not_include_preexisting_errors
+        doc = Nokogiri::HTML::Document.parse "<html><body><div></div><jimmy></jimmy></body></html>"
+        assert doc.errors.any?{|err| err.to_s =~ /jimmy/}, "assert on setup"
+
+        context_node = doc.at_css "div"
+        frag = Nokogiri::HTML::DocumentFragment.new doc, "<hello>oh, hello there.</hello>", context_node
+        assert frag.errors.any?{|err| err.to_s =~ /Tag hello invalid/}, "errors should be on the context node's document"
+        assert frag.errors.none?{|err| err.to_s =~ /jimmy/}, "errors should not include pre-existing document errors"
+      end
     end
   end
 end

data/test/html/test_node.rb

@@ -127,6 +127,15 @@ module Nokogiri
         assert_equal 'foo&bar&baz', node['href']
       end
 
+      def test_parse_config_option
+        node = @html.at('div')
+        options = nil
+        node.parse("<div></div>") do |config|
+          options = config
+        end
+        assert_equal Nokogiri::XML::ParseOptions::DEFAULT_HTML, options.to_i
+      end
+
       def test_fragment_handler_does_not_regurge_on_invalid_attributes
         iframe = %Q{<iframe style="width: 0%; height: 0px" src="http://someurl" allowtransparency></iframe>}
         assert_nothing_raised { @html.at('div').fragment(iframe) }

data/test/xml/sax/test_parser.rb

@@ -24,7 +24,7 @@ module Nokogiri
 
           assert block_called
 
-          assert_equal ['a', '&b'], doc.start_elements.first.last
+          assert_equal [['foo', [['a', '&b']]]], doc.start_elements
         end
 
         def test_parser_context_yielded_in_memory
@@ -40,7 +40,7 @@ module Nokogiri
 
           assert block_called
 
-          assert_equal ['a', '&b'], doc.start_elements.first.last
+          assert_equal [['foo', [['a', '&b']]]], doc.start_elements
         end
 
         def test_xml_decl
@@ -296,7 +296,15 @@ module Nokogiri
           @parser.parse_memory(<<-eoxml)
             <p id="asdfasdf">Paragraph 1</p>
           eoxml
-          assert_equal [["p", ["id", "asdfasdf"]]],
+          assert_equal [["p", [["id", "asdfasdf"]]]],
+                       @parser.document.start_elements
+        end
+
+        def test_start_element_attrs_include_namespaces
+          @parser.parse_memory(<<-eoxml)
+            <p xmlns:foo='http://foo.example.com/'>Paragraph 1</p>
+          eoxml
+          assert_equal [["p", [['xmlns:foo', 'http://foo.example.com/']]]],
                        @parser.document.start_elements
         end
 

data/test/xml/sax/test_parser_context.rb

@@ -6,6 +6,56 @@ module Nokogiri
   module XML
     module SAX
       class TestParserContext < Nokogiri::SAX::TestCase
+        def setup
+          @xml = '<hello>
+
+world
+<inter>
+    <net>
+    </net>
+</inter>
+
+</hello>'
+        end
+
+        class Counter < Nokogiri::XML::SAX::Document
+          attr_accessor :context, :lines, :columns
+          def initialize
+            @context = nil
+            @lines   = []
+            @columns = []
+          end
+
+          def start_element name, attrs = []
+            @lines << [name, context.line]
+            @columns << [name, context.column]
+          end
+        end
+
+        def test_line_numbers
+          sax_handler = Counter.new
+
+          parser = Nokogiri::XML::SAX::Parser.new(sax_handler)
+          parser.parse(@xml) do |ctx|
+            sax_handler.context = ctx
+          end
+
+          assert_equal [["hello", 1], ["inter", 4], ["net", 5]],
+            sax_handler.lines
+        end
+
+        def test_column_numbers
+          sax_handler = Counter.new
+
+          parser = Nokogiri::XML::SAX::Parser.new(sax_handler)
+          parser.parse(@xml) do |ctx|
+            sax_handler.context = ctx
+          end
+
+          assert_equal [["hello", 7], ["inter", 7], ["net", 9]],
+            sax_handler.columns
+        end
+
         def test_replace_entities
           pc = ParserContext.new StringIO.new('<root />'), 'UTF-8'
           pc.replace_entities = false