nokogiri 1.0.0-x86-mswin32-60

data/lib/nokogiri/xml/node_set.rb

@@ -0,0 +1,183 @@
+module Nokogiri
+  module XML
+    class NodeSet
+      include Enumerable
+
+      attr_accessor :document
+
+      def initialize
+        yield self if block_given?
+      end
+
+      ###
+      # Get the first element of the NodeSet.
+      def first
+        self[0]
+      end
+
+      ###
+      # Get the last element of the NodeSet.
+      def last
+        self[length - 1]
+      end
+
+      ###
+      # Is this NodeSet empty?
+      def empty?
+        length == 0
+      end
+
+      ###
+      # Insert +datum+ before the first Node in this NodeSet
+      def before datum
+        first.before datum
+      end
+
+      ###
+      # Insert +datum+ after the last Node in this NodeSet
+      def after datum
+        last.after datum
+      end
+
+      ###
+      # Append +node+ to the NodeSet.
+      def << node
+        push(node)
+      end
+
+      ###
+      # Unlink this NodeSet and all Node objects it contains from their
+      # current context.
+      def unlink
+        each { |node| node.unlink }
+        self.document = nil
+        self
+      end
+      alias :remove :unlink
+
+      ###
+      # Search this document for +paths+
+      def search *paths
+        sub_set = NodeSet.new
+        document.decorate(sub_set)
+        each do |node|
+          node.search(*paths).each do |sub_node|
+            sub_set << sub_node
+          end
+        end
+        sub_set.document = document
+        sub_set
+      end
+      alias :/ :search
+      alias :xpath :search
+      alias :css :search
+
+      ###
+      # If path is a string, search this document for +path+ returning the
+      # first Node.  Otherwise, index in to the array with +path+.
+      def at path, ns = {}
+        return self[path] if path.is_a?(Numeric)
+        search(path, ns).first
+      end
+
+      ###
+      # Append the class attribute +name+ to all Node objects in the NodeSet.
+      def add_class name
+        each do |el|
+          next unless el.respond_to? :get_attribute
+          classes = el.get_attribute('class').to_s.split(" ")
+          el.set_attribute('class', classes.push(name).uniq.join(" "))
+        end
+        self
+      end
+
+      ###
+      # Remove the class attribute +name+ from all Node objects in the NodeSet.
+      def remove_class name = nil
+        each do |el|
+          next unless el.respond_to? :get_attribute
+          if name
+            classes = el.get_attribute('class').to_s.split(" ")
+            el.set_attribute('class', (classes - [name]).uniq.join(" "))
+          else
+            el.remove_attribute("class")
+          end
+        end
+        self
+      end
+
+      ###
+      # Set the attribute +key+ to +value+ or the return value of +blk+
+      # on all Node objects in the NodeSet.
+      def attr key, value = nil, &blk
+        if value or blk
+          each do |el|
+            el.set_attribute(key, value || blk[el])
+          end
+          return self      
+        end    
+        if key.is_a? Hash
+          key.each { |k,v| self.attr(k,v) }
+          return self
+        else
+          return self[0].get_attribute(key)
+        end
+      end
+      alias_method :set, :attr
+
+      ###
+      # Remove the attributed named +name+ from all Node objects in the NodeSet
+      def remove_attr name
+        each do |el|
+          next unless el.respond_to? :remove_attribute
+          el.remove_attribute(name)
+        end
+        self      
+      end
+
+      ###
+      # Iterate over each node, yielding  to +block+
+      def each(&block)
+        x = 0
+        while x < length
+          yield self[x]
+          x += 1
+        end
+      end
+
+      ###
+      # Get the inner text of all contained Node objects
+      def inner_text
+        collect{|j| j.inner_text}.join('')
+      end
+      alias :text :inner_text
+
+      ###
+      # Wrap this NodeSet with +html+ or the results of the builder in +blk+
+      def wrap(html, &blk)
+        each do |j|
+          new_parent = Nokogiri.make(html, &blk)
+          j.replace(new_parent)
+          nest = new_parent
+          if nest.child
+            nest = nest.child until nest.child.nil?
+          end
+          j.parent = nest
+        end
+        self
+      end
+
+      def to_s
+        map { |x| x.to_s }.join
+      end
+
+      def to_html
+        map { |x| x.to_html }.join('')
+      end
+
+      def size
+        length
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/notation.rb

@@ -0,0 +1,6 @@
+module Nokogiri
+  module XML
+    class Notation < Struct.new(:name, :public_id, :system_id)
+    end
+  end
+end

data/lib/nokogiri/xml/reader.rb

@@ -0,0 +1,14 @@
+module Nokogiri
+  module XML
+    class Reader
+      include Enumerable
+
+      def each(&block)
+        while node = self.read
+          block.call(node)
+        end
+      end
+      private :initialize
+    end
+  end
+end

data/lib/nokogiri/xml/sax.rb

@@ -0,0 +1,9 @@
+require 'nokogiri/xml/sax/document'
+require 'nokogiri/xml/sax/parser'
+
+module Nokogiri
+  module XML
+    module SAX
+    end
+  end
+end

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

@@ -0,0 +1,59 @@
+module Nokogiri
+  module XML
+    module SAX
+      class Document
+        ###
+        # Called when document starts parsing
+        def start_document
+        end
+
+        ###
+        # Called when document ends parsing
+        def end_document
+        end
+
+        ###
+        # Called at the beginning of an element
+        # +name+ is the name of the tag with +attrs+ as attributes
+        def start_element name, attrs = []
+        end
+
+        ###
+        # Called at the end of an element
+        # +name+ is the tag name
+        def end_element name
+        end
+
+        ###
+        # Characters read between a tag
+        # +string+ contains the character data
+        def characters string
+        end
+
+        ###
+        # Called when comments are encountered
+        # +string+ contains the comment data
+        def comment string
+        end
+
+        ###
+        # Called on document warnings
+        # +string+ contains the warning
+        def warning string
+        end
+
+        ###
+        # Called on document errors
+        # +string+ contains the error
+        def error string
+        end
+
+        ###
+        # Called when cdata blocks are found
+        # +string+ contains the cdata content
+        def cdata_block string
+        end
+      end
+    end
+  end
+end

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

@@ -0,0 +1,33 @@
+module Nokogiri
+  module XML
+    module SAX
+      class Parser
+        attr_accessor :document
+        def initialize(doc = XML::SAX::Document.new)
+          @document = doc
+        end
+
+        ###
+        # Parse given +thing+ which may be a string containing xml, or an
+        # IO object.
+        def parse thing
+          parse_memory(thing.is_a?(IO) ? thing.read : thing)
+        end
+
+        ###
+        # Parse given +io+
+        def parse_io io
+          parse_memory io.read
+        end
+
+        ###
+        # Parse a file with +filename+
+        def parse_file filename
+          raise Errno::ENOENT unless File.exists?(filename)
+          raise Errno::EISDIR if File.directory?(filename)
+          native_parse_file filename
+        end
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/syntax_error.rb

@@ -0,0 +1,21 @@
+module Nokogiri
+  module XML
+    class SyntaxError < SyntaxError
+      def none?
+        level == 0
+      end
+
+      def warning?
+        level == 1
+      end
+
+      def error?
+        level == 2
+      end
+
+      def fatal?
+        level == 3
+      end
+    end
+  end
+end

data/lib/nokogiri/xml/text.rb

@@ -0,0 +1,6 @@
+module Nokogiri
+  module XML
+    class Text < Node
+    end
+  end
+end

data/lib/nokogiri/xml/xpath.rb

@@ -0,0 +1,6 @@
+module Nokogiri
+  module XML
+    class XPath
+    end
+  end
+end

data/lib/nokogiri/xml/xpath_context.rb

@@ -0,0 +1,14 @@
+module Nokogiri
+  module XML
+    class XPathContext
+
+      def register_namespaces(namespaces)
+        namespaces.each do |k, v|
+          k = k.gsub(/.*:/,'') # strip off 'xmlns:' or 'xml:'
+          register_ns(k, v)
+        end
+      end
+
+    end
+  end
+end

data/lib/nokogiri/xslt.rb

@@ -0,0 +1,11 @@
+require 'nokogiri/xslt/stylesheet'
+
+module Nokogiri
+  module XSLT
+    class << self
+      def parse(string)
+        Stylesheet.parse_stylesheet_doc(XML.parse(string))
+      end
+    end
+  end
+end

data/lib/nokogiri/xslt/stylesheet.rb

@@ -0,0 +1,6 @@
+module Nokogiri
+  module XSLT
+    class Stylesheet
+    end
+  end
+end

data/nokogiri.gemspec

@@ -0,0 +1,34 @@
+Gem::Specification.new do |s|
+  s.name = %q{nokogiri}
+  s.version = "1.0.0"
+  s.platform = %q{x86-mswin32-60}
+
+  s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
+  s.authors = ["Aaron Patterson"]
+  s.date = %q{2008-10-30}
+  s.description = %q{Nokogiri (鋸) is an HTML, XML, SAX, and Reader parser.}
+  s.email = ["aaronp@rubyforge.org"]
+  s.extra_rdoc_files = ["History.txt", "Manifest.txt", "README.ja.txt", "README.txt"]
+  s.files = ["History.txt", "Manifest.txt", "README.ja.txt", "README.txt", "Rakefile", "ext/nokogiri/extconf.rb", "ext/nokogiri/html_document.c", "ext/nokogiri/html_document.h", "ext/nokogiri/html_sax_parser.c", "ext/nokogiri/html_sax_parser.h", "ext/nokogiri/native.c", "ext/nokogiri/native.h", "ext/nokogiri/xml_cdata.c", "ext/nokogiri/xml_cdata.h", "ext/nokogiri/xml_document.c", "ext/nokogiri/xml_document.h", "ext/nokogiri/xml_dtd.c", "ext/nokogiri/xml_dtd.h", "ext/nokogiri/xml_node.c", "ext/nokogiri/xml_node.h", "ext/nokogiri/xml_node_set.c", "ext/nokogiri/xml_node_set.h", "ext/nokogiri/xml_reader.c", "ext/nokogiri/xml_reader.h", "ext/nokogiri/xml_sax_parser.c", "ext/nokogiri/xml_sax_parser.h", "ext/nokogiri/xml_syntax_error.c", "ext/nokogiri/xml_syntax_error.h", "ext/nokogiri/xml_text.c", "ext/nokogiri/xml_text.h", "ext/nokogiri/xml_xpath.c", "ext/nokogiri/xml_xpath.h", "ext/nokogiri/xml_xpath_context.c", "ext/nokogiri/xml_xpath_context.h", "ext/nokogiri/xslt_stylesheet.c", "ext/nokogiri/xslt_stylesheet.h", "lib/nokogiri.rb", "lib/nokogiri/css.rb", "lib/nokogiri/css/generated_parser.rb", "lib/nokogiri/css/generated_tokenizer.rb", "lib/nokogiri/css/node.rb", "lib/nokogiri/css/parser.rb", "lib/nokogiri/css/parser.y", "lib/nokogiri/css/tokenizer.rb", "lib/nokogiri/css/tokenizer.rex", "lib/nokogiri/css/xpath_visitor.rb", "lib/nokogiri/decorators.rb", "lib/nokogiri/decorators/hpricot.rb", "lib/nokogiri/decorators/hpricot/node.rb", "lib/nokogiri/decorators/hpricot/node_set.rb", "lib/nokogiri/decorators/hpricot/xpath_visitor.rb", "lib/nokogiri/hpricot.rb", "lib/nokogiri/html.rb", "lib/nokogiri/html/builder.rb", "lib/nokogiri/html/document.rb", "lib/nokogiri/html/sax/parser.rb", "lib/nokogiri/version.rb", "lib/nokogiri/xml.rb", "lib/nokogiri/xml/after_handler.rb", "lib/nokogiri/xml/before_handler.rb", "lib/nokogiri/xml/builder.rb", "lib/nokogiri/xml/cdata.rb", "lib/nokogiri/xml/document.rb", "lib/nokogiri/xml/dtd.rb", "lib/nokogiri/xml/element.rb", "lib/nokogiri/xml/entity_declaration.rb", "lib/nokogiri/xml/node.rb", "lib/nokogiri/xml/node_set.rb", "lib/nokogiri/xml/notation.rb", "lib/nokogiri/xml/reader.rb", "lib/nokogiri/xml/sax.rb", "lib/nokogiri/xml/sax/document.rb", "lib/nokogiri/xml/sax/parser.rb", "lib/nokogiri/xml/syntax_error.rb", "lib/nokogiri/xml/text.rb", "lib/nokogiri/xml/xpath.rb", "lib/nokogiri/xml/xpath_context.rb", "lib/nokogiri/xslt.rb", "lib/nokogiri/xslt/stylesheet.rb", "nokogiri.gemspec", "test/css/test_nthiness.rb", "test/css/test_parser.rb", "test/css/test_tokenizer.rb", "test/css/test_xpath_visitor.rb", "test/files/staff.xml", "test/files/staff.xslt", "test/files/tlm.html", "test/helper.rb", "test/hpricot/files/basic.xhtml", "test/hpricot/files/boingboing.html", "test/hpricot/files/cy0.html", "test/hpricot/files/immob.html", "test/hpricot/files/pace_application.html", "test/hpricot/files/tenderlove.html", "test/hpricot/files/uswebgen.html", "test/hpricot/files/utf8.html", "test/hpricot/files/week9.html", "test/hpricot/files/why.xml", "test/hpricot/load_files.rb", "test/hpricot/test_alter.rb", "test/hpricot/test_builder.rb", "test/hpricot/test_parser.rb", "test/hpricot/test_paths.rb", "test/hpricot/test_preserved.rb", "test/hpricot/test_xml.rb", "test/html/sax/test_parser.rb", "test/html/test_builder.rb", "test/html/test_document.rb", "test/test_convert_xpath.rb", "test/test_nokogiri.rb", "test/test_reader.rb", "test/test_xslt_transforms.rb", "test/xml/sax/test_parser.rb", "test/xml/test_builder.rb", "test/xml/test_cdata.rb", "test/xml/test_document.rb", "test/xml/test_dtd.rb", "test/xml/test_node.rb", "test/xml/test_node_set.rb", "test/xml/test_text.rb", "ext/nokogiri/iconv.dll", "ext/nokogiri/libexslt.dll", "ext/nokogiri/libxml2.dll", "ext/nokogiri/libxslt.dll", "ext/nokogiri/zlib1.dll", "ext/nokogiri/native.so"]
+  s.has_rdoc = true
+  s.homepage = %q{http://nokogiri.rubyforge.org/}
+  s.rdoc_options = ["--main", "README.txt"]
+  s.require_paths = ["lib", "ext"]
+  s.rubyforge_project = %q{nokogiri}
+  s.rubygems_version = %q{1.2.0}
+  s.summary = %q{Nokogiri (鋸) is an HTML, XML, SAX, and Reader parser.}
+  s.test_files = ["test/css/test_nthiness.rb", "test/css/test_parser.rb", "test/css/test_tokenizer.rb", "test/css/test_xpath_visitor.rb", "test/hpricot/test_alter.rb", "test/hpricot/test_builder.rb", "test/hpricot/test_parser.rb", "test/hpricot/test_paths.rb", "test/hpricot/test_preserved.rb", "test/hpricot/test_xml.rb", "test/html/sax/test_parser.rb", "test/html/test_builder.rb", "test/html/test_document.rb", "test/test_convert_xpath.rb", "test/test_nokogiri.rb", "test/test_reader.rb", "test/test_xslt_transforms.rb", "test/xml/sax/test_parser.rb", "test/xml/test_builder.rb", "test/xml/test_cdata.rb", "test/xml/test_document.rb", "test/xml/test_dtd.rb", "test/xml/test_node.rb", "test/xml/test_node_set.rb", "test/xml/test_text.rb"]
+
+  if s.respond_to? :specification_version then
+    current_version = Gem::Specification::CURRENT_SPECIFICATION_VERSION
+    s.specification_version = 2
+
+    if current_version >= 3 then
+      s.add_development_dependency(%q<hoe>, [">= 1.7.0"])
+    else
+      s.add_dependency(%q<hoe>, [">= 1.7.0"])
+    end
+  else
+    s.add_dependency(%q<hoe>, [">= 1.7.0"])
+  end
+end

data/test/css/test_nthiness.rb

@@ -0,0 +1,159 @@
+require File.expand_path(File.join(File.dirname(__FILE__), '..', "helper"))
+
+module Nokogiri
+  module CSS
+    class TestNthiness < Nokogiri::TestCase
+      def setup
+        doc = <<EOF
+<html>
+<table>
+  <tr><td>row1 </td></tr>
+  <tr><td>row2 </td></tr>
+  <tr><td>row3 </td></tr>
+  <tr><td>row4 </td></tr>
+  <tr><td>row5 </td></tr>
+  <tr><td>row6 </td></tr>
+  <tr><td>row7 </td></tr>
+  <tr><td>row8 </td></tr>
+  <tr><td>row9 </td></tr>
+  <tr><td>row10 </td></tr>
+  <tr><td>row11 </td></tr>
+  <tr><td>row12 </td></tr>
+  <tr><td>row13 </td></tr>
+  <tr><td>row14 </td></tr>
+</table>
+<div>
+  <b>bold1 </b>
+  <i>italic1 </i>
+  <b>bold2 </b>
+  <i>italic2 </i>
+  <p>para1 </p>
+  <b>bold3 </b>
+</div>
+<div>
+  <p>para2 </p>
+  <p>para3 </p>
+</div>
+<div>
+  <p>para4 </p>
+</div>
+<p class='empty'></p>
+<p class='not-empty'><b></b></p>
+</html>
+EOF
+        @parser = Nokogiri.Hpricot doc
+      end
+
+
+      def test_even
+        assert_result_rows [2,4,6,8,10,12,14], @parser.search("table/tr:nth(even)")
+      end
+
+      def test_odd
+        assert_result_rows [1,3,5,7,9,11,13], @parser.search("table/tr:nth(odd)")
+      end
+
+      def test_2n
+        assert_equal @parser.search("table/tr:nth(even)").inner_text, @parser.search("table/tr:nth(2n)").inner_text
+      end
+
+      def test_2np1
+        assert_equal @parser.search("table/tr:nth(odd)").inner_text, @parser.search("table/tr:nth(2n+1)").inner_text
+      end
+
+      def test_4np3
+        assert_result_rows [3,7,11], @parser.search("table/tr:nth(4n+3)")
+      end
+
+      def test_3np4
+        assert_result_rows [4,7,10,13], @parser.search("table/tr:nth(3n+4)")
+      end
+
+      def test_mnp3
+        assert_result_rows [1,2,3], @parser.search("table/tr:nth(-n+3)")
+      end
+
+      def test_np3
+        assert_result_rows [3,4,5,6,7,8,9,10,11,12,13,14], @parser.search("table/tr:nth(n+3)")
+      end
+
+      def test_first
+        assert_result_rows [1], @parser.search("table/tr:first")
+        assert_result_rows [1], @parser.search("table/tr:first()")
+      end
+
+      def test_last
+        assert_result_rows [14], @parser.search("table/tr:last")
+        assert_result_rows [14], @parser.search("table/tr:last()")
+      end
+
+      def test_first_child
+        assert_result_rows [1], @parser.search("div/b:first-child"), "bold"
+        assert_result_rows [1], @parser.search("table/tr:first-child")
+      end
+
+      def test_last_child
+        assert_result_rows [3], @parser.search("div/b:last-child"), "bold"
+        assert_result_rows [14], @parser.search("table/tr:last-child")
+      end
+
+      def test_first_of_type
+        assert_result_rows [1], @parser.search("table/tr:first-of-type")
+        assert_result_rows [1], @parser.search("div/b:first-of-type"), "bold"
+      end
+
+      def test_last_of_type
+        assert_result_rows [14], @parser.search("table/tr:last-of-type")
+        assert_result_rows [3], @parser.search("div/b:last-of-type"), "bold"
+      end
+
+      def test_only_of_type
+        assert_result_rows [1,4], @parser.search("div/p:only-of-type"), "para"
+      end
+
+      def test_only_child
+        assert_result_rows [4], @parser.search("div/p:only-child"), "para"
+      end
+
+      def test_empty
+        result = @parser.search("p:empty")
+        assert_equal 1, result.size, "unexpected number of rows returned: '#{result.inner_text}'"
+        assert_equal 'empty', result.first['class']
+      end
+
+      def test_parent
+        result = @parser.search("p:parent")
+        assert_equal 5, result.size
+        0.upto(3) do |j|
+          assert_equal "para#{j+1} ", result[j].inner_text
+        end
+        assert_equal "not-empty", result[4]['class']
+      end
+
+      def test_siblings
+        doc = <<-EOF
+<html><body><div>
+<p id="1">p1 </p>
+<p id="2">p2 </p>
+<p id="3">p3 </p>
+<p id="4">p4 </p>
+<p id="5">p5 </p>
+EOF
+        parser = Nokogiri.Hpricot doc
+        
+        assert_equal 2, parser.search("#3 ~ p").size
+        assert_equal "p4 p5 ", parser.search("#3 ~ p").inner_text
+        assert_equal 0, parser.search("#5 ~ p").size
+
+        assert_equal 1, parser.search("#3 + p").size
+        assert_equal "p4 ", parser.search("#3 + p").inner_text
+        assert_equal 0, parser.search("#5 + p").size
+      end
+
+      def assert_result_rows intarray, result, word="row"
+        assert_equal intarray.size, result.size, "unexpected number of rows returned: '#{result.inner_text}'"
+        assert_equal intarray.map{|j| "#{word}#{j}"}.join(' '), result.inner_text.strip, result.inner_text
+      end
+    end
+  end
+end