nokogiri 1.6.8.rc3 → 1.6.8

data/ext/nokogiri/xml_reader.c

@@ -219,19 +219,6 @@ static VALUE reader_attribute(VALUE self, VALUE name)
   name = StringValue(name) ;
 
   value = xmlTextReaderGetAttribute(reader, (xmlChar*)StringValueCStr(name));
-  if(value == NULL) {
-    /* this section is an attempt to workaround older versions of libxml that
-       don't handle namespaces properly in all attribute-and-friends functions */
-    xmlChar *prefix = NULL ;
-    xmlChar *localname = xmlSplitQName2((xmlChar*)StringValueCStr(name), &prefix);
-    if (localname != NULL) {
-      value = xmlTextReaderLookupNamespace(reader, localname);
-      xmlFree(localname) ;
-    } else {
-      value = xmlTextReaderLookupNamespace(reader, prefix);
-    }
-    xmlFree(prefix);
-  }
   if(value == NULL) return Qnil;
 
   rb_value = NOKOGIRI_STR_NEW2(value);

data/ext/nokogiri/xml_sax_parser.c

@@ -18,16 +18,16 @@ static void start_document(void * ctx)
 
   if(NULL != ctxt && ctxt->html != 1) {
     if(ctxt->standalone != -1) {  /* -1 means there was no declaration */
-      VALUE encoding = ctxt->encoding ?
-        NOKOGIRI_STR_NEW2(ctxt->encoding) :
-        Qnil;
+      VALUE encoding = Qnil ;
+      if (ctxt->encoding) {
+        encoding = NOKOGIRI_STR_NEW2(ctxt->encoding) ;
+      } else if (ctxt->input && ctxt->input->encoding) {
+        encoding = NOKOGIRI_STR_NEW2(ctxt->input->encoding) ;
+      }
 
-      VALUE version = ctxt->version ?
-        NOKOGIRI_STR_NEW2(ctxt->version) :
-        Qnil;
+      VALUE version = ctxt->version ? NOKOGIRI_STR_NEW2(ctxt->version) : Qnil;
 
       VALUE standalone = Qnil;
-
       switch(ctxt->standalone)
       {
         case 0:

data/lib/nokogiri.rb

@@ -114,15 +114,19 @@ module Nokogiri
     def Slop(*args, &block)
       Nokogiri(*args, &block).slop!
     end
+
+    def install_default_aliases
+      # Make sure to support some popular encoding aliases not known by
+      # all iconv implementations.
+      {
+        'Windows-31J' => 'CP932',	# Windows-31J is the IANA registered name of CP932.
+      }.each { |alias_name, name|
+        EncodingHandler.alias(name, alias_name) if EncodingHandler[alias_name].nil?
+      }
+    end
   end
 
-  # Make sure to support some popular encoding aliases not known by
-  # all iconv implementations.
-  {
-    'Windows-31J' => 'CP932',	# Windows-31J is the IANA registered name of CP932.
-  }.each { |alias_name, name|
-    EncodingHandler.alias(name, alias_name) if EncodingHandler[alias_name].nil?
-  }
+  Nokogiri.install_default_aliases
 end
 
 ###

data/lib/nokogiri/html/document.rb

@@ -173,7 +173,7 @@ module Nokogiri
 
           if string_or_io.respond_to?(:read)
             url ||= string_or_io.respond_to?(:path) ? string_or_io.path : nil
-            if !encoding
+            unless encoding
               # Libxml2's parser has poor support for encoding
               # detection.  First, it does not recognize the HTML5
               # style meta charset declaration.  Secondly, even if it
@@ -196,7 +196,9 @@ module Nokogiri
           end
 
           # read_memory pukes on empty docs
-          return new if string_or_io.nil? or string_or_io.empty?
+          if string_or_io.nil? or string_or_io.empty?
+            return encoding ? new.tap { |i| i.encoding = encoding } : new
+          end
 
           encoding ||= EncodingReader.detect_encoding(string_or_io)
 

data/lib/nokogiri/version.rb

@@ -1,6 +1,6 @@
 module Nokogiri
   # The version of Nokogiri you are using
-  VERSION = '1.6.8.rc3'
+  VERSION = '1.6.8'
 
   class VersionInfo # :nodoc:
     def jruby?

data/lib/nokogiri/xml/document.rb

@@ -49,7 +49,7 @@ module Nokogiri
           if options.strict?
             raise Nokogiri::XML::SyntaxError.new("Empty document")
           else
-            return new
+            return encoding ? new.tap { |i| i.encoding = encoding } : new
           end
         end
 

data/lib/nokogiri/xml/parse_options.rb

@@ -2,6 +2,28 @@ module Nokogiri
   module XML
     ###
     # Parse options for passing to Nokogiri.XML or Nokogiri.HTML
+    #
+    # == Building combinations of parse options
+    # You can build your own combinations of these parse options by using any of the following methods:
+    # *Note*: All examples attempt to set the +RECOVER+ & +NOENT+ options. All examples use Ruby 2 optional parameter syntax.
+    # [Ruby's bitwise operators] You can use the Ruby bitwise operators to set various combinations.
+    #   <code>Nokogiri.XML('<content>Chapter 1</content', options: Nokogiri::XML::ParseOptions.new((1 << 0) | (1 << 1)))</code>
+    # [Method chaining] Every option has an equivalent method in lowercase. You can chain these methods together to set various combinations.
+    #   <code>Nokogiri.XML('<content>Chapter 1</content', options: Nokogiri::XML::ParseOptions.new.recover.noent)</code>
+    # [Using Ruby Blocks] You can also setup parse combinations in the block passed to Nokogiri.XML or Nokogiri.HTML
+    #   <code>Nokogiri.XML('<content>Chapter 1</content') {|config| config.recover.noent}</code>
+    #
+    # == Removing particular parse options
+    # You can also remove options from an instance of +ParseOptions+ dynamically.
+    # Every option has an equivalent <code>no{option}</code> method in lowercase. You can call these methods on an instance of +ParseOptions+ to remove the option.
+    # Note that this is not available for +STRICT+.
+    #
+    #   # Setting the RECOVER & NOENT options...
+    #   options = Nokogiri::XML::ParseOptions.new.recover.noent
+    #   # later...
+    #   options.norecover # Removes the Nokogiri::XML::ParseOptions::RECOVER option
+    #   options.nonoent # Removes the Nokogiri::XML::ParseOptions::NOENT option
+    #
     class ParseOptions
       # Strict parsing
       STRICT      = 0

data/tasks/test.rb

@@ -5,6 +5,11 @@ namespace :test do
     Rake::Task["test"].invoke
   end
 
+  task :installed do
+    ENV['RUBY_FLAGS'] = "-w -Itest:."
+    sh 'rake test'
+  end
+
   desc "find call-seq in the rdoc"
   task :rdoc_call_seq => 'docs' do
     Dir['doc/**/*.html'].each { |docfile|

data/test/html/test_document.rb

@@ -670,6 +670,32 @@ eohtml
         Nokogiri::XML::Element.new("div", doc).set_attribute('id', 'unique-issue-1262')
         assert_equal 0, doc.errors.length
       end
+
+      it "skips encoding for script tags" do
+        html = Nokogiri::HTML <<-EOHTML
+        <html>
+          <head>
+            <script>var isGreater = 4 > 5;</script>
+          </head>
+          <body></body>
+        </html>
+        EOHTML
+        node = html.xpath("//script").first
+        assert_equal("var isGreater = 4 > 5;", node.inner_html)
+      end
+
+      it "skips encoding for style tags" do
+        html = Nokogiri::HTML <<-EOHTML
+        <html>
+          <head>
+            <style>tr > div { display:block; }</style>
+          </head>
+          <body></body>
+        </html>
+        EOHTML
+        node = html.xpath("//style").first
+        assert_equal("tr > div { display:block; }", node.inner_html)
+      end
     end
   end
 end

data/test/html/test_document_encoding.rb

@@ -80,6 +80,11 @@ module Nokogiri
         assert_equal ['http://tenderlovemaking.com/'],
           doc.css('a').map { |a| a['href'] }
       end
+
+      def test_empty_doc_encoding
+        encoding = 'US-ASCII'
+        assert_equal encoding, Nokogiri::HTML.parse(nil, nil, encoding).encoding
+      end
     end
 
     class TestDocumentEncodingDetection < Nokogiri::TestCase

data/test/html/test_document_fragment.rb

@@ -21,6 +21,11 @@ module Nokogiri
         assert_equal 'EUC-JP', f.document.encoding
         assert_equal "こんにちは！", f.content
       end
+      
+      def test_unlink_empty_document
+        frag = Nokogiri::HTML::DocumentFragment.parse('').unlink # must_not_raise
+        assert_nil frag.parent
+      end
 
       def test_colons_are_not_removed
         doc = Nokogiri::HTML::DocumentFragment.parse("<span>3:30pm</span>")

data/test/test_encoding_handler.rb

@@ -5,6 +5,8 @@ require "helper"
 class TestEncodingHandler < Nokogiri::TestCase
   def teardown
     Nokogiri::EncodingHandler.clear_aliases!
+    #Replace default aliases removed by clear_aliases!
+    Nokogiri.install_default_aliases
   end
 
   def test_get

data/test/test_xslt_transforms.rb

@@ -35,6 +35,39 @@ class TestXsltTransforms < Nokogiri::TestCase
     assert_match %r{<h1></h1>|<h1/>}, result
   end
 
+  def test_xml_declaration
+    input_xml = <<-EOS
+<?xml version="1.0" encoding="utf-8"?>
+<report>
+  <title>My Report</title>
+</report>
+EOS
+
+    input_xsl = <<-EOS
+<?xml version="1.0" encoding="utf-8"?>
+<xsl:stylesheet version="1.0" xmlns:xsl="http://www.w3.org/1999/XSL/Transform">
+  <xsl:output method="xml" version="1.0" encoding="utf-8" indent="yes"/>
+  <xsl:template match="/">
+    <html>
+      <head>
+        <title><xsl:value-of select="report/title"/></title>
+      </head>
+      <body>
+        <h1><xsl:value-of select="report/title"/></h1>
+      </body>
+    </html>
+  </xsl:template>
+</xsl:stylesheet>
+EOS
+
+    require 'nokogiri'
+
+    xml = ::Nokogiri::XML(input_xml)
+    xsl = ::Nokogiri::XSLT(input_xsl)
+
+    assert_includes xsl.apply_to(xml), '<?xml version="1.0" encoding="utf-8"?>'
+  end
+
   def test_transform_with_output_style
     xslt = ""
     if Nokogiri.jruby?

data/test/xml/sax/test_parser.rb

@@ -44,13 +44,21 @@ module Nokogiri
         end
 
         def test_xml_decl
-          {
-            ''          => nil,
-            '<?xml version="1.0" ?>'                  => ['1.0'],
-            '<?xml version="1.0" encoding="UTF-8" ?>' => ['1.0', 'UTF-8'],
-            '<?xml version="1.0" standalone="yes"?>'  => ['1.0', 'yes'],
-            '<?xml version="1.0" standalone="no"?>'   => ['1.0', 'no'],
-          }.each do |decl,value|
+          [
+            ['', nil],
+            ['<?xml version="1.0" ?>',
+              ['1.0']],
+            ['<?xml version="1.0" encoding="UTF-8" ?>',
+              ['1.0', 'UTF-8']],
+            ['<?xml version="1.0" standalone="yes"?>',
+              ['1.0', 'yes']],
+            ['<?xml version="1.0" standalone="no"?>',
+              ['1.0', 'no']],
+            ['<?xml version="1.0" encoding="UTF-8" standalone="no"?>',
+              ['1.0', "UTF-8", 'no']],
+            ['<?xml version="1.0" encoding="ISO-8859-1" standalone="yes"?>',
+              ['1.0', "ISO-8859-1", 'yes']]
+          ].each do |decl, value|
             parser = XML::SAX::Parser.new(Doc.new)
 
             xml = "#{decl}\n<root />"

data/test/xml/test_document_encoding.rb

@@ -21,6 +21,11 @@ module Nokogiri
           assert_equal 'UTF-8', Nokogiri::LIBXML_VERSION.encoding.name
         end
       end
+
+      def test_empty_doc_encoding
+        encoding = 'US-ASCII'
+        assert_equal encoding, Nokogiri::XML(nil, nil, encoding).encoding
+      end
     end
   end
 end

data/test/xml/test_document_fragment.rb

@@ -224,6 +224,18 @@ module Nokogiri
         Nokogiri::XML::Comment.new(frag,'moo')
       end
 
+      def test_issue_1077_parsing_of_frozen_strings
+        input = <<-EOS
+<?xml version="1.0" encoding="utf-8"?>
+<library>
+  <book title="I like turtles"/>
+</library>
+EOS
+        input.freeze
+
+        Nokogiri::XML::DocumentFragment.parse(input) # assert_nothing_raised
+      end
+
       if Nokogiri.uses_libxml?
         def test_for_libxml_in_context_fragment_parsing_bug_workaround
           10.times do

data/test/xml/test_node_attributes.rb

@@ -26,6 +26,12 @@ module Nokogiri
         assert_equal nil, node['lang']
       end
 
+      def test_unknown_namespace_prefix_should_not_be_removed
+        doc = Nokogiri::XML ''
+        elem = doc.create_element 'foo', 'bar:attr' => 'something'
+        assert_equal elem.attribute_nodes.first.name, 'bar:attr'
+      end
+
       def test_set_prefixed_attributes
         doc = Nokogiri::XML %Q{<root xmlns:foo="x"/>}
 

data/test/xml/test_node_reparenting.rb

@@ -49,6 +49,15 @@ module Nokogiri
         #   end
         # end
 
+
+        before do
+          @doc  = Nokogiri::XML "<root><a1>First node</a1><a2>Second node</a2><a3>Third <bx />node</a3></root>"
+          @doc2 = @doc.dup
+          @fragment_string = "<b1>foo</b1><b2>bar</b2>"
+          @fragment        = Nokogiri::XML::DocumentFragment.parse @fragment_string
+          @node_set        = Nokogiri::XML("<root><b1>foo</b1><b2>bar</b2></root>").xpath("/root/node()")
+        end
+
         {
           :add_child            => {:target => "/root/a1",        :returns_self => false, :children_tags => %w[text b1 b2]},
           :<<                   => {:target => "/root/a1",        :returns_self => true, :children_tags => %w[text b1 b2]},
@@ -67,15 +76,6 @@ module Nokogiri
           :next=                => {:target => "/root/a1/text()", :returns_self => false, :children_tags => %w[text b1 b2]},
           :after                => {:target => "/root/a1/text()", :returns_self => true,  :children_tags => %w[text b1 b2]}
         }.each do |method, params|
-
-          before do
-            @doc  = Nokogiri::XML "<root><a1>First node</a1><a2>Second node</a2><a3>Third <bx />node</a3></root>"
-            @doc2 = @doc.dup
-            @fragment_string = "<b1>foo</b1><b2>bar</b2>"
-            @fragment        = Nokogiri::XML::DocumentFragment.parse @fragment_string
-            @node_set        = Nokogiri::XML("<root><b1>foo</b1><b2>bar</b2></root>").xpath("/root/node()")
-          end
-
           describe "##{method}" do
             describe "passed a Node" do
               [:current, :another].each do |which|
@@ -215,7 +215,7 @@ module Nokogiri
             end
           end
 
-          describe "given a parent node with a non-default namespace" do
+          describe "given a parent node with a default and non-default namespace" do
             before do
               @doc = Nokogiri::XML(<<-eoxml)
                 <root xmlns="http://tenderlovemaking.com/" xmlns:foo="http://flavorjon.es/">
@@ -223,18 +223,156 @@ module Nokogiri
                   </first>
                 </root>
               eoxml
+              assert @node = @doc.at('//xmlns:first')
+              @child = Nokogiri::XML::Node.new('second', @doc)
             end
 
-            describe "and a child node with a namespace matching the parent's non-default namespace" do
-              it "inserts a node that inherits the matching parent namespace" do
-                assert node = @doc.at('//xmlns:first')
-                child = Nokogiri::XML::Node.new('second', @doc)
+            describe "and a child with a namespace matching the parent's default namespace" do
+              describe "and as the default prefix" do
+                before do
+                  @ns = @child.add_namespace(nil, 'http://tenderlovemaking.com/')
+                  @child.namespace = @ns
+                end
 
-                ns = @doc.root.namespace_definitions.detect { |x| x.prefix == "foo" }
-                child.namespace = ns
+                it "inserts a node that inherits the parent's default namespace" do
+                  @node.add_child(@child)
+                  assert reparented = @doc.at('//bar:second', "bar" => "http://tenderlovemaking.com/")
+                  assert reparented.namespace_definitions.empty?
+                  assert_equal @ns, reparented.namespace
+                  assert_equal(
+                    {
+                      "xmlns"     => "http://tenderlovemaking.com/",
+                      "xmlns:foo" => "http://flavorjon.es/",
+                    },
+                    reparented.namespaces)
+                end
+              end
+
+              describe "but with a different prefix" do
+                before do
+                  @ns = @child.add_namespace("baz", 'http://tenderlovemaking.com/')
+                  @child.namespace = @ns
+                end
 
-                node.add_child(child)
-                assert @doc.at('//foo:second', "foo" => "http://flavorjon.es/")
+                it "inserts a node that uses its own namespace" do
+                  @node.add_child(@child)
+                  assert reparented = @doc.at('//bar:second', "bar" => "http://tenderlovemaking.com/")
+                  assert reparented.namespace_definitions.include?(@ns)
+                  assert_equal @ns, reparented.namespace
+                  assert_equal(
+                    {
+                      "xmlns"     => "http://tenderlovemaking.com/",
+                      "xmlns:foo" => "http://flavorjon.es/",
+                      "xmlns:baz" => "http://tenderlovemaking.com/",
+                    },
+                    reparented.namespaces)
+                end
+              end
+            end
+
+            describe "and a child with a namespace matching the parent's non-default namespace" do
+              describe "set by #namespace=" do
+                before do
+                  @ns = @doc.root.namespace_definitions.detect { |x| x.prefix == "foo" }
+                  @child.namespace = @ns
+                end
+
+                it "inserts a node that inherits the matching parent namespace" do
+                  @node.add_child(@child)
+                  assert reparented = @doc.at('//bar:second', "bar" => "http://flavorjon.es/")
+                  assert reparented.namespace_definitions.empty?
+                  assert_equal @ns, reparented.namespace
+                  assert_equal(
+                    {
+                      "xmlns"     => "http://tenderlovemaking.com/",
+                      "xmlns:foo" => "http://flavorjon.es/",
+                    },
+                    reparented.namespaces)
+                end
+              end
+
+              describe "with the same prefix" do
+                before do
+                  @ns = @child.add_namespace("foo", 'http://flavorjon.es/')
+                  @child.namespace = @ns
+                end
+
+                it "inserts a node that uses the parent's namespace" do
+                  @node.add_child(@child)
+                  assert reparented = @doc.at('//bar:second', "bar" => "http://flavorjon.es/")
+                  assert reparented.namespace_definitions.empty?
+                  assert_equal @ns, reparented.namespace
+                  assert_equal(
+                    {
+                      "xmlns"     => "http://tenderlovemaking.com/",
+                      "xmlns:foo" => "http://flavorjon.es/",
+                    },
+                    reparented.namespaces)
+                end
+              end
+
+              describe "as the default prefix" do
+                before do
+                  @ns = @child.add_namespace(nil, 'http://flavorjon.es/')
+                  @child.namespace = @ns
+                end
+
+                it "inserts a node that keeps its namespace" do
+                  @node.add_child(@child)
+                  assert reparented = @doc.at('//bar:second', "bar" => "http://flavorjon.es/")
+                  assert reparented.namespace_definitions.include?(@ns)
+                  assert_equal @ns, reparented.namespace
+                  assert_equal(
+                    {
+                      "xmlns"     => "http://flavorjon.es/",
+                      "xmlns:foo" => "http://flavorjon.es/",
+                    },
+                    reparented.namespaces)
+                end
+              end
+
+              describe "but with a different prefix" do
+                before do
+                  @ns = @child.add_namespace('baz', 'http://flavorjon.es/')
+                  @child.namespace = @ns
+                end
+
+                it "inserts a node that keeps its namespace" do
+                  @node.add_child(@child)
+                  assert reparented = @doc.at('//bar:second', "bar" => "http://flavorjon.es/")
+                  assert reparented.namespace_definitions.include?(@ns)
+                  assert_equal @ns, reparented.namespace
+                  assert_equal(
+                    {
+                      "xmlns"     =>"http://tenderlovemaking.com/",
+                      "xmlns:foo" =>"http://flavorjon.es/",
+                      "xmlns:baz" =>"http://flavorjon.es/",
+                    },
+                    reparented.namespaces)
+                end
+              end
+            end
+
+            describe "and a child node with a default namespace not matching the parent's default namespace and a namespace matching a parent namespace but with a different prefix" do
+              before do
+                @ns = @child.add_namespace(nil, 'http://example.org/')
+                @child.namespace = @ns
+                @ns2 = @child.add_namespace('baz', 'http://tenderlovemaking.com/')
+              end
+
+              it "inserts a node that keeps its namespace" do
+                @node.add_child(@child)
+                assert reparented = @doc.at('//bar:second', "bar" => "http://example.org/")
+                assert reparented.namespace_definitions.include?(@ns)
+                assert reparented.namespace_definitions.include?(@ns2)
+                assert_equal @ns, reparented.namespace
+                assert_equal(
+                  {
+                    "xmlns"     => "http://example.org/",
+                    "xmlns:foo" => "http://flavorjon.es/",
+                    "xmlns:baz" => "http://tenderlovemaking.com/",
+                  },
+                  reparented.namespaces)
               end
             end
           end
@@ -368,6 +506,43 @@ module Nokogiri
             end
           end
         end
+
+        describe "reparenting into another document" do
+          it "correctly sets default namespace of a reparented node" do
+            # issue described in #391
+            # thanks to Nick Canzoneri @nickcanz for this test case!
+            source_doc = Nokogiri::XML <<-EOX
+<?xml version="1.0" encoding="utf-8"?>
+<Wix xmlns="http://schemas.microsoft.com/wix/2006/wi">
+    <Product>
+        <Package />
+        <Directory Id="TARGETDIR" Name="SourceDir">
+            <Component>
+                <File />
+            </Component>
+        </Directory>
+    </Product>
+</Wix>
+EOX
+
+            dest_doc = Nokogiri::XML <<-EOX
+<?xml version="1.0" encoding="utf-8"?>
+<Wix xmlns='http://schemas.microsoft.com/wix/2006/wi'>
+  <Fragment Id='MSIComponents'>
+      <DirectoryRef Id='InstallDir'>
+      </DirectoryRef>
+  </Fragment>
+</Wix>
+EOX
+
+            stuff = source_doc.at_css("Directory[Id='TARGETDIR']")
+            insert_point = dest_doc.at_css("DirectoryRef[Id='InstallDir']")
+            insert_point.children = stuff.children()
+
+            assert_no_match(/default:/, insert_point.children.to_xml)
+            assert_match(/<Component>/, insert_point.children.to_xml)
+          end
+        end
       end
     end
   end