nokogiri 1.5.2-java → 1.5.3-java

data/lib/nokogiri/css/parser.y

@@ -9,12 +9,13 @@ rule
     : selector COMMA simple_selector_1toN {
         result = [val.first, val.last].flatten
       }
+    | prefixless_combinator_selector { result = val.flatten }
     | simple_selector_1toN { result = val.flatten }
     ;
   combinator
     : PLUS { result = :DIRECT_ADJACENT_SELECTOR }
     | GREATER { result = :CHILD_SELECTOR }
-    | TILDE { result = :PRECEDING_SELECTOR }
+    | TILDE { result = :FOLLOWING_SELECTOR }
     | S { result = :DESCENDANT_SELECTOR }
     | DOUBLESLASH { result = :DESCENDANT_SELECTOR }
     | SLASH { result = :CHILD_SELECTOR }
@@ -59,6 +60,11 @@ rule
         )
       }
     ;
+  prefixless_combinator_selector
+    : combinator simple_selector_1toN {
+        result = Node.new(val.first, [nil, val.last])
+      }
+    ;
   simple_selector_1toN
     : simple_selector combinator simple_selector_1toN {
         result = Node.new(val[1], [val.first, val.last])
@@ -88,7 +94,7 @@ rule
     |
     ;
   attrib
-    : LSQUARE namespaced_ident attrib_val_0or1 RSQUARE {
+    : LSQUARE attrib_name attrib_val_0or1 RSQUARE {
         result = Node.new(:ATTRIBUTE_CONDITION,
           [val[1]] + (val[2] || [])
         )
@@ -105,6 +111,18 @@ rule
         )
       }
     ;
+  attrib_name
+    : namespace '|' IDENT {
+        result = Node.new(:ELEMENT_NAME,
+          [[val.first, val.last].compact.join(':')]
+        )
+      }
+    | IDENT {
+        # Default namespace is not applied to attributes.
+        # So we don't add prefix "xmlns:" as in namespaced_ident.
+        result = Node.new(:ELEMENT_NAME, [val.first])
+      }
+    ;
   function
     : FUNCTION RPAREN {
         result = Node.new(:FUNCTION, [val.first.strip])

data/lib/nokogiri/css/xpath_visitor.rb

@@ -126,13 +126,13 @@ module Nokogiri
       {
         'combinator'                => ' and ',
         'direct_adjacent_selector'  => "/following-sibling::*[1]/self::",
-        'preceding_selector'        => "/following-sibling::",
+        'following_selector'        => "/following-sibling::",
         'descendant_selector'       => '//',
         'child_selector'            => '/',
       }.each do |k,v|
         class_eval %{
           def visit_#{k} node
-            "\#{node.value.first.accept(self)}#{v}\#{node.value.last.accept(self)}"
+            "\#{node.value.first.accept(self) if node.value.first}#{v}\#{node.value.last.accept(self)}"
           end
         }
       end

data/lib/nokogiri/html/document.rb

@@ -6,7 +6,8 @@ module Nokogiri
       # then nil is returned.
       def meta_encoding
         meta = meta_content_type and
-          /charset\s*=\s*([\w-]+)/i.match(meta['content'])[1]
+          match = /charset\s*=\s*([\w-]+)/i.match(meta['content']) and
+          match[1]
       end
 
       ###

data/lib/nokogiri/html/element_description_defaults.rb

@@ -3,7 +3,7 @@ module Nokogiri
     class ElementDescription
 
       # Methods are defined protected by method_defined? because at
-      # this point the C-library or Java library is alraedy loaded,
+      # this point the C-library or Java library is already loaded,
       # and we don't want to clobber any methods that have been
       # defined there.
 

data/lib/nokogiri/version.rb

@@ -1,6 +1,6 @@
 module Nokogiri
   # The version of Nokogiri you are using
-  VERSION = '1.5.2'
+  VERSION = '1.5.3'
 
   class VersionInfo # :nodoc:
     def jruby?

data/lib/nokogiri/xml/document.rb

@@ -132,7 +132,7 @@ module Nokogiri
       # 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
+      # traverses the entire graph, and also has to bring each node across the
       # libxml bridge into a ruby object.
       def collect_namespaces
         ns = {}

data/lib/nokogiri/xml/parse_options.rb

@@ -25,11 +25,11 @@ module Nokogiri
       NOBLANKS    = 1 << 8
       # use the SAX1 interface internally
       SAX1        = 1 << 9
-      # Implement XInclude substitition
+      # Implement XInclude substitution
       XINCLUDE    = 1 << 10
       # Forbid network access
       NONET       = 1 << 11
-      # Do not reuse the context dictionnary
+      # Do not reuse the context dictionary
       NODICT      = 1 << 12
       # remove redundant namespaces declarations
       NSCLEAN     = 1 << 13

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

@@ -33,7 +33,7 @@ module Nokogiri
     #   parser = Nokogiri::XML::SAX::Parser.new(MyDocument.new)
     #
     #   # Feed the parser some XML
-    #   parser.parse(File.read(ARGV[0], 'rb'))
+    #   parser.parse(File.open(ARGV[0]))
     #
     # Now my document handler will be called when each node starts, and when
     # then document ends.  To see what kinds of events are available, take

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

@@ -25,7 +25,7 @@ module Nokogiri
       #   parser = Nokogiri::XML::SAX::Parser.new(MyDoc.new)
       #
       #   # Send some XML to the parser
-      #   parser.parse(File.read(ARGV[0]))
+      #   parser.parse(File.open(ARGV[0]))
       #
       # For more information about SAX parsers, see Nokogiri::XML::SAX.  Also
       # see Nokogiri::XML::SAX::Document for the available events.

data/lib/nokogiri/xslt.rb

@@ -15,7 +15,7 @@ module Nokogiri
   end
 
   ###
-  # See Nokogiri::XSLT::Stylesheet for creating and maniuplating
+  # See Nokogiri::XSLT::Stylesheet for creating and manipulating
   # Stylesheet object.
   module XSLT
     class << self

data/test/css/test_parser.rb

@@ -6,6 +6,10 @@ module Nokogiri
       def setup
         super
         @parser = Nokogiri::CSS::Parser.new
+        @parser_with_ns = Nokogiri::CSS::Parser.new({
+          "xmlns" => "http://default.example.com/",
+          "hoge" => "http://hoge.example.com/",
+        })
       end
 
       def test_extra_single_quote
@@ -214,6 +218,33 @@ module Nokogiri
                       @parser.parse("E + F G")
       end
 
+      def test_child_selector
+        assert_xpath("//a//b/i", @parser.parse('a b>i'))
+        assert_xpath("//a//b/i", @parser.parse('a b > i'))
+        assert_xpath("//a/b/i", @parser.parse('a > b > i'))
+      end
+
+      def test_prefixless_child_selector
+        assert_xpath("./a", @parser.parse('>a'))
+        assert_xpath("./a", @parser.parse('> a'))
+        assert_xpath("./a//b/i", @parser.parse('>a b>i'))
+        assert_xpath("./a/b/i", @parser.parse('> a > b > i'))
+      end
+
+      def test_prefixless_preceding_sibling_selector
+        assert_xpath("./following-sibling::a", @parser.parse('~a'))
+        assert_xpath("./following-sibling::a", @parser.parse('~ a'))
+        assert_xpath("./following-sibling::a//b/following-sibling::i", @parser.parse('~a b~i'))
+        assert_xpath("./following-sibling::a//b/following-sibling::i", @parser.parse('~ a b ~ i'))
+      end
+
+      def test_prefixless_direct_adjacent_selector
+        assert_xpath("./following-sibling::*[1]/self::a", @parser.parse('+a'))
+        assert_xpath("./following-sibling::*[1]/self::a", @parser.parse('+ a'))
+        assert_xpath("./following-sibling::*[1]/self::a/following-sibling::*[1]/self::b", @parser.parse('+a+b'))
+        assert_xpath("./following-sibling::*[1]/self::a/following-sibling::*[1]/self::b", @parser.parse('+ a + b'))
+      end
+
       def test_attribute
         assert_xpath  "//h1[@a = 'Tender Lovemaking']",
                       @parser.parse("h1[a='Tender Lovemaking']")
@@ -292,6 +323,13 @@ module Nokogiri
         # assert_xpath ['//x/y', '//y/z'], @parser.parse('x > y | y > z')
       end
 
+      def test_attributes_with_namespace
+        ## Default namespace is not applied to attributes.
+        ## So this must be @class, not @xmlns:class.
+        assert_xpath "//xmlns:a[@class = 'bar']", @parser_with_ns.parse("a[class='bar']")
+        assert_xpath "//xmlns:a[@hoge:class = 'bar']", @parser_with_ns.parse("a[hoge|class='bar']")
+      end
+
       def assert_xpath expecteds, asts
         expecteds = [expecteds].flatten
         expecteds.zip(asts).each do |expected, actual|

data/test/files/to_be_xincluded.xml

@@ -0,0 +1,2 @@
+<?xml version="1.0" encoding="utf-8"?>
+<included>this snippet is to be included from xinclude.xml</included>

data/test/files/xinclude.xml

@@ -0,0 +1,4 @@
+<?xml version="1.0" encoding="utf-8"?>
+<test xmlns:xi="http://www.w3.org/2001/XInclude">
+    <xi:include href="to_be_xincluded.xml"/>
+</test>

data/test/helper.rb

@@ -11,25 +11,25 @@ warn "#{__FILE__}:#{__LINE__}: version info: #{Nokogiri::VERSION_INFO.inspect}"
 
 module Nokogiri
   class TestCase < MiniTest::Spec
-    ASSETS_DIR      = File.expand_path File.join(File.dirname(__FILE__), 'files')
-    XML_FILE        = File.join(ASSETS_DIR, 'staff.xml')
-    XSLT_FILE       = File.join(ASSETS_DIR, 'staff.xslt')
-    EXSLT_FILE      = File.join(ASSETS_DIR, 'exslt.xslt')
-    EXML_FILE       = File.join(ASSETS_DIR, 'exslt.xml')
-    HTML_FILE       = File.join(ASSETS_DIR, 'tlm.html')
-    NICH_FILE       = File.join(ASSETS_DIR, '2ch.html')
-    SHIFT_JIS_XML   = File.join(ASSETS_DIR, 'shift_jis.xml')
-    SHIFT_JIS_HTML  = File.join(ASSETS_DIR, 'shift_jis.html')
-    ENCODING_XHTML_FILE = File.join(ASSETS_DIR, 'encoding.xhtml')
-    ENCODING_HTML_FILE  = File.join(ASSETS_DIR, 'encoding.html')
-    NOENCODING_FILE  = File.join(ASSETS_DIR, 'noencoding.html')
-    METACHARSET_FILE  = File.join(ASSETS_DIR, 'metacharset.html')
-    PO_XML_FILE     = File.join(ASSETS_DIR, 'po.xml')
-    PO_SCHEMA_FILE  = File.join(ASSETS_DIR, 'po.xsd')
+    ASSETS_DIR          = File.expand_path File.join(File.dirname(__FILE__), 'files')
     ADDRESS_SCHEMA_FILE = File.join(ASSETS_DIR, 'address_book.rlx')
-    ADDRESS_XML_FILE = File.join(ASSETS_DIR, 'address_book.xml')
-    SNUGGLES_FILE   = File.join(ASSETS_DIR, 'snuggles.xml')
-    XML_XINCLUDE_FILE = File.join(ASSETS_DIR, 'xinclude.xml')
+    ADDRESS_XML_FILE    = File.join(ASSETS_DIR, 'address_book.xml')
+    ENCODING_HTML_FILE  = File.join(ASSETS_DIR, 'encoding.html')
+    ENCODING_XHTML_FILE = File.join(ASSETS_DIR, 'encoding.xhtml')
+    EXML_FILE           = File.join(ASSETS_DIR, 'exslt.xml')
+    EXSLT_FILE          = File.join(ASSETS_DIR, 'exslt.xslt')
+    HTML_FILE           = File.join(ASSETS_DIR, 'tlm.html')
+    METACHARSET_FILE    = File.join(ASSETS_DIR, 'metacharset.html')
+    NICH_FILE           = File.join(ASSETS_DIR, '2ch.html')
+    NOENCODING_FILE     = File.join(ASSETS_DIR, 'noencoding.html')
+    PO_SCHEMA_FILE      = File.join(ASSETS_DIR, 'po.xsd')
+    PO_XML_FILE         = File.join(ASSETS_DIR, 'po.xml')
+    SHIFT_JIS_HTML      = File.join(ASSETS_DIR, 'shift_jis.html')
+    SHIFT_JIS_XML       = File.join(ASSETS_DIR, 'shift_jis.xml')
+    SNUGGLES_FILE       = File.join(ASSETS_DIR, 'snuggles.xml')
+    XML_FILE            = File.join(ASSETS_DIR, 'staff.xml')
+    XML_XINCLUDE_FILE   = File.join(ASSETS_DIR, 'xinclude.xml')
+    XSLT_FILE           = File.join(ASSETS_DIR, 'staff.xslt')
 
     def teardown
       if ENV['NOKOGIRI_GC']
@@ -68,33 +68,6 @@ module Nokogiri
     alias :assert_not_nil       :refute_nil
     alias :assert_raise         :assert_raises
     alias :assert_not_equal     :refute_equal
-
-    def assert_nothing_raised(*args)
-      self._assertions += 1
-      if Module === args.last
-        msg = nil
-      else
-        msg = args.pop
-      end
-      begin
-        line = __LINE__; yield
-      rescue Exception => e
-        bt = e.backtrace
-        as = e.instance_of?(MiniTest::Assertion)
-        if as
-          ans = /\A#{Regexp.quote(__FILE__)}:#{line}:in /o
-          bt.reject! {|ln| ans =~ ln}
-        end
-        if ((args.empty? && !as) ||
-            args.any? {|a| a.instance_of?(Module) ? e.is_a?(a) : e.class == a })
-          msg = message(msg) { "Exception raised:\n<#{mu_pp(e)}>" }
-          raise MiniTest::Assertion, msg.call, bt
-        else
-          raise
-        end
-      end
-      nil
-    end
   end
 
   module SAX

data/test/html/sax/test_parser.rb

@@ -12,19 +12,21 @@ module Nokogiri
 
         def test_parse_empty_document
           # This caused a segfault in libxml 2.6.x
-          assert_nothing_raised { @parser.parse '' }
+          assert_nil @parser.parse ''
         end
 
         def test_parse_empty_file
           # Make sure empty files don't break stuff
           empty_file_name =  File.join(Dir.tmpdir, 'bogus.xml')
           FileUtils.touch empty_file_name
-          assert_nothing_raised { @parser.parse_file empty_file_name }
+          # assert_nothing_raised do
+            @parser.parse_file empty_file_name
+          # end
         end
 
         def test_parse_file
           @parser.parse_file(HTML_FILE)
-          
+
           # Take a look at the comment in test_parse_document to know
           # a possible reason to this difference.
           if Nokogiri.uses_libxml?

data/test/html/sax/test_parser_context.rb

@@ -7,15 +7,13 @@ module Nokogiri
     module SAX
       class TestParserContext < Nokogiri::SAX::TestCase
         def test_from_io
-          assert_nothing_raised do
-            ParserContext.new StringIO.new('fo'), 'UTF-8'
-          end
+          ctx = ParserContext.new StringIO.new('fo'), 'UTF-8'
+          assert ctx
         end
 
         def test_from_string
-          assert_nothing_raised do
-            ParserContext.new 'blah blah'
-          end
+          ctx = ParserContext.new 'blah blah'
+          assert ctx
         end
 
         def test_parse_with
@@ -26,20 +24,20 @@ module Nokogiri
         end
 
         def test_parse_with_sax_parser
-          assert_nothing_raised do
+          # assert_nothing_raised do
             xml = "<root />"
             ctx = ParserContext.new xml
             parser = Parser.new Doc.new
             ctx.parse_with parser
-          end
+          # end
         end
 
         def test_from_file
-          assert_nothing_raised do
+          # assert_nothing_raised do
             ctx = ParserContext.file HTML_FILE, 'UTF-8'
             parser = Parser.new Doc.new
             ctx.parse_with parser
-          end
+          # end
         end
       end
     end

data/test/html/test_document.rb

@@ -125,8 +125,10 @@ module Nokogiri
 
       def test_meta_encoding
         assert_equal 'UTF-8', @html.meta_encoding
+      end
 
-        html = Nokogiri::HTML(<<-eohtml)
+      def test_meta_encoding_is_strict_about_http_equiv
+        doc = Nokogiri::HTML(<<-eohtml)
 <html>
   <head>
     <meta http-equiv="X-Content-Type" content="text/html; charset=Shift_JIS">
@@ -136,7 +138,21 @@ module Nokogiri
   </body>
 </html>
         eohtml
-        assert_nil html.meta_encoding
+        assert_nil doc.meta_encoding
+      end
+
+      def test_meta_encoding_handles_malformed_content_charset
+        doc = Nokogiri::HTML(<<EOHTML)
+<html>
+  <head>
+    <meta http-equiv="Content-type" content="text/html; utf-8" />
+  </head>
+  <body>
+    foo
+  </body>
+</html>
+EOHTML
+        assert_nil doc.meta_encoding
       end
 
       def test_meta_encoding=
@@ -232,9 +248,7 @@ eohtml
       end
 
       def test_parse_handles_nil_gracefully
-        assert_nothing_raised do
-          @doc = Nokogiri::HTML::Document.parse(nil)
-        end
+        @doc = Nokogiri::HTML::Document.parse(nil)
         assert_instance_of Nokogiri::HTML::Document, @doc
       end
 

data/test/html/test_node.rb

@@ -101,9 +101,7 @@ module Nokogiri
         assert another_node, 'should have a node'
 
         # This used to segv
-        assert_nothing_raised do
-          node.add_previous_sibling another_node
-        end
+        assert node.add_previous_sibling another_node
       end
 
       def test_swap
@@ -138,7 +136,7 @@ module Nokogiri
 
       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) }
+        assert @html.at('div').fragment(iframe)
       end
 
       def test_fragment

data/test/test_reader.rb

@@ -422,4 +422,67 @@ class TestReader < Nokogiri::TestCase
     end
   end
 
+  def test_correct_outer_xml_inclusion
+    xml = Nokogiri::XML::Reader.from_io(StringIO.new(<<-eoxml))
+      <root-element>
+        <children>
+          <child n="1">
+            <field>child-1</field>
+          </child>
+          <child n="2">
+            <field>child-2</field>
+          </child>
+          <child n="3">
+            <field>child-3</field>
+          </child>
+        </children>
+      </root-element>
+    eoxml
+
+    nodelengths = []
+    has_child2 = []
+
+    xml.each do |node|
+      if node.node_type == Nokogiri::XML::Reader::TYPE_ELEMENT and node.name == "child"
+        nodelengths << node.outer_xml.length
+        has_child2 << !!(node.outer_xml =~ /child-2/)
+      end
+    end
+
+    assert_equal(nodelengths[0], nodelengths[1])
+    assert(has_child2[1])
+    assert(!has_child2[0])
+  end
+
+  def test_correct_inner_xml_inclusion
+    xml = Nokogiri::XML::Reader.from_io(StringIO.new(<<-eoxml))
+      <root-element>
+        <children>
+          <child n="1">
+            <field>child-1</field>
+          </child>
+          <child n="2">
+            <field>child-2</field>
+          </child>
+          <child n="3">
+            <field>child-3</field>
+          </child>
+        </children>
+      </root-element>
+    eoxml
+
+    nodelengths = []
+    has_child2 = []
+
+    xml.each do |node|
+      if node.node_type == Nokogiri::XML::Reader::TYPE_ELEMENT and node.name == "child"
+        nodelengths << node.inner_xml.length
+        has_child2 << !!(node.inner_xml =~ /child-2/)
+      end
+    end
+
+    assert_equal(nodelengths[0], nodelengths[1])
+    assert(has_child2[1])
+    assert(!has_child2[0])
+  end
 end