RubyGems - nokogiri - Versions diffs - 1.7.2 → 1.8.0 - Mend

nokogiri 1.7.2 → 1.8.0

Potentially problematic release.

This version of nokogiri might be problematic. Click here for more details.

Files changed (50) hide show

checksums.yaml +4 -4
data/.cross_rubies +4 -4
data/.travis.yml +43 -24
data/CHANGELOG.md +54 -6
data/Gemfile +8 -7
data/Gemfile-libxml-ruby +3 -0
data/LICENSE-DEPENDENCIES.md +1612 -0
data/{LICENSE.txt → LICENSE.md} +1 -1
data/Manifest.txt +5 -8
data/README.md +8 -5
data/Rakefile +15 -31
data/appveyor.yml +2 -0
data/dependencies.yml +12 -7
data/ext/nokogiri/extconf.rb +12 -17
data/ext/nokogiri/nokogiri.h +0 -10
data/ext/nokogiri/xml_attr.c +12 -8
data/ext/nokogiri/xml_node.c +17 -14
data/ext/nokogiri/xml_sax_push_parser.c +56 -12
data/lib/nokogiri/html/sax/parser.rb +10 -0
data/lib/nokogiri/version.rb +5 -4
data/lib/nokogiri/xml/document.rb +9 -9
data/lib/nokogiri/xml/node.rb +7 -7
data/lib/nokogiri/xml/node_set.rb +12 -7
data/lib/nokogiri/xml/sax/parser.rb +6 -7
data/lib/nokogiri/xml/searchable.rb +34 -25
data/lib/nokogiri/xml/syntax_error.rb +24 -1
data/test/decorators/test_slop.rb +4 -1
data/test/helper.rb +10 -0
data/test/html/sax/test_parser.rb +27 -0
data/test/html/test_document.rb +12 -1
data/test/html/test_document_encoding.rb +1 -3
data/test/html/test_document_fragment.rb +3 -0
data/test/xml/sax/test_push_parser.rb +48 -0
data/test/xml/test_attr.rb +7 -0
data/test/xml/test_document.rb +1 -1
data/test/xml/test_document_fragment.rb +27 -0
data/test/xml/test_entity_reference.rb +2 -2
data/test/xml/test_node.rb +12 -15
data/test/xml/test_node_reparenting.rb +14 -0
data/test/xml/test_node_set.rb +8 -6
data/test/xml/test_reader.rb +19 -0
data/test/xml/test_syntax_error.rb +21 -15
data/test/xml/test_unparented_node.rb +54 -11
data/test/xml/test_xpath.rb +23 -6
metadata +31 -19
data/suppressions/nokogiri_ree-1.8.7.358.supp +0 -61
data/suppressions/nokogiri_ruby-1.8.7.370.supp +0 -0
data/suppressions/nokogiri_ruby-1.9.2.320.supp +0 -28
data/suppressions/nokogiri_ruby-1.9.3.327.supp +0 -28
data/test_all +0 -105

data/test/helper.rb CHANGED

@@ -7,7 +7,17 @@ require 'tempfile'
 require 'pp'
 require 'nokogiri'
 if ENV['TEST_NOKOGIRI_WITH_LIBXML_RUBY']
+  #
+  #  if you'd like to test with the libxml-ruby gem loaded, it's
+  #  recommended that you set
+  #
+  #    BUNDLE_GEMFILE=Gemfile-libxml-ruby
+  #
+  #  which will a) bundle that gem, and b) set the appropriate env var to
+  #  trigger this block
+  #
   require 'libxml'
   warn "#{__FILE__}:#{__LINE__}: loaded libxml-ruby '#{LibXML::XML::VERSION}'"
 end

data/test/html/sax/test_parser.rb CHANGED

@@ -132,6 +132,33 @@ module Nokogiri
           ], @parser.document.start_elements
         end
+        HTML_WITH_BR_TAG = <<-EOF
+        <html>
+          <head></head>
+          <body>
+            <div>
+              hello
+              <br>
+            </div>
+            <div>
+              hello again
+            </div>
+          </body>
+        </html>
+        EOF
+        def test_parsing_dom_error_from_string
+          @parser.parse(HTML_WITH_BR_TAG)
+          assert_equal 6, @parser.document.start_elements.length
+        end
+        def test_parsing_dom_error_from_io
+          @parser.parse(StringIO.new(HTML_WITH_BR_TAG))
+          assert_equal 6, @parser.document.start_elements.length
+        end
         def test_empty_processing_instruction
           @parser.parse_memory("<strong>this will segfault<?strong>")
         end

data/test/html/test_document.rb CHANGED

@@ -15,7 +15,7 @@ module Nokogiri
       end
       def test_does_not_fail_with_illformatted_html
-        doc = Nokogiri::HTML('"</html>";'.force_encoding(Encoding::BINARY))
+        doc = Nokogiri::HTML('"</html>";'.dup.force_encoding(Encoding::BINARY))
         assert_not_nil doc
       end
@@ -696,6 +696,17 @@ eohtml
         node = html.xpath("//style").first
         assert_equal("tr > div { display:block; }", node.inner_html)
       end
+      it "does not fail when converting to_html using explicit encoding" do
+        html_fragment=<<-eos
+  <img width="16" height="16" src="images/icon.gif" border="0" alt="Inactive hide details for &quot;User&quot; ---19/05/2015 12:55:29---Provvediamo subito nell&#8217;integrare">
+        eos
+        doc = Nokogiri::HTML(html_fragment, nil, 'ISO-8859-1')
+        html = doc.to_html
+        assert html.index("src=\"images/icon.gif\"")
+        assert_equal 'ISO-8859-1', html.encoding.name
+      end
     end
   end
 end

data/test/html/test_document_encoding.rb CHANGED

@@ -128,9 +128,7 @@ module Nokogiri
           assert_equal(title, doc_from_string_enc.at('//title/text()').text)
           assert_equal(title, doc_from_string.at('//title/text()').text)
           assert_equal(title, doc_from_file_enc.at('//title/text()').text)
-          unless Nokogiri.jruby? && file == ENCODING_HTML_FILE
-            assert_equal(title, doc_from_file.at('//title/text()').text)
-          end
+          assert_equal(title, doc_from_file.at('//title/text()').text)
           evil = (0..72).map { |i| '超' * i + '悪い事を構想中。' }

data/test/html/test_document_fragment.rb CHANGED

@@ -234,6 +234,9 @@ module Nokogiri
       end
       def test_element_children_counts
+        if Nokogiri.uses_libxml? && Nokogiri::VERSION_INFO['libxml']['loaded'] <= "2.9.1"
+          skip "#elements doesn't work in 2.9.1, see 1793a5a for history"
+        end
         doc = Nokogiri::HTML::DocumentFragment.parse("   <div>  </div>\n   ")
         assert_equal 1, doc.element_children.count
       end

data/test/xml/sax/test_push_parser.rb CHANGED

@@ -151,6 +151,54 @@ module Nokogiri
           assert_equal "Gau\337", @parser.document.data.join
           assert_equal [["r"]], @parser.document.end_elements
         end
+        def test_replace_entities_attribute_behavior
+          if Nokogiri.uses_libxml?
+            # initially false
+            assert_equal false, @parser.replace_entities
+            # can be set to true
+            @parser.replace_entities = true
+            assert_equal true, @parser.replace_entities
+            # can be set to false
+            @parser.replace_entities = false
+            assert_equal false, @parser.replace_entities
+          else
+            # initially true
+            assert_equal true, @parser.replace_entities
+            # ignore attempts to set to false
+            @parser.replace_entities = false # TODO: should we raise an exception here?
+            assert_equal true, @parser.replace_entities
+          end
+        end
+        def test_untouched_entities
+          skip("entities are always replaced in pure Java version") if Nokogiri.jruby?
+          @parser.<<(<<-eoxml)
+            <p id="asdf&amp;asdf">
+              <!-- This is a comment -->
+              Paragraph 1 &amp; 2
+            </p>
+          eoxml
+          @parser.finish
+          assert_equal [["p", [["id", "asdf&#38;asdf"]]]], @parser.document.start_elements
+          assert_equal "Paragraph 1 & 2", @parser.document.data.join.strip
+        end
+        def test_replaced_entities
+          @parser.replace_entities = true
+          @parser.<<(<<-eoxml)
+            <p id="asdf&amp;asdf">
+              <!-- This is a comment -->
+              Paragraph 1 &amp; 2
+            </p>
+          eoxml
+          @parser.finish
+          assert_equal [["p", [["id", "asdf&asdf"]]]], @parser.document.start_elements
+          assert_equal "Paragraph 1 & 2", @parser.document.data.join.strip
+        end
       end
     end
   end

data/test/xml/test_attr.rb CHANGED

@@ -11,6 +11,13 @@ module Nokogiri
         }
       end
+      def test_new_raises_argerror_on_nondocument
+        document = Nokogiri::XML "<root><foo/></root>"
+        assert_raises ArgumentError do
+          Nokogiri::XML::Attr.new document.at_css("foo"), "bar"
+        end
+      end
       def test_content=
         xml = Nokogiri::XML.parse(File.read(XML_FILE), XML_FILE)
         address = xml.xpath('//address')[3]

data/test/xml/test_document.rb CHANGED

@@ -221,7 +221,7 @@ module Nokogiri
       end
       def test_pp
-        out = StringIO.new('')
+        out = StringIO.new(String.new)
         ::PP.pp @xml, out
         assert_operator out.string.length, :>, 0
       end

data/test/xml/test_document_fragment.rb CHANGED

@@ -136,6 +136,23 @@ module Nokogiri
         assert_equal expected, search_xpath
       end
+      def test_fragment_css_search_with_whitespace_and_node_removal
+        # The same xml without leading whitespace in front of the first line
+        # does not expose the error. Putting both nodes on the same line
+        # instead also fixes the crash.
+        fragment = Nokogiri::XML::DocumentFragment.parse <<-EOXML
+          <p id="content">hi</p> x <!--y--> <p>another paragraph</p>
+        EOXML
+        children = fragment.css('p')
+        assert_equal 2, children.length
+        # removing the last node instead does not yield the error. Probably the
+        # node removal leaves around two consecutive text nodes which make the
+        # css search crash?
+        children.first.remove
+        assert_equal 1, fragment.xpath('.//p | self::p').length
+        assert_equal 1, fragment.css('p').length
+      end
       def test_fragment_search_three_ways
         frag = Nokogiri::XML::Document.new.fragment '<p id="content">foo</p><p id="content">bar</p>'
         expected = frag.xpath('./*[@id = "content"]')
@@ -219,6 +236,16 @@ module Nokogiri
         assert fragment.children.respond_to?(:awesome!), fragment.children.class
       end
+      def test_decorator_is_applied_to_empty_set
+        x = Module.new do
+          def awesome!
+          end
+        end
+        util_decorate(@xml, x)
+        fragment = Nokogiri::XML::DocumentFragment.new(@xml, "")
+        assert fragment.children.respond_to?(:awesome!), fragment.children.class
+      end
       def test_add_node_to_doc_fragment_segfault
         frag = Nokogiri::XML::DocumentFragment.new(@xml, '<p>hello world</p>')
         Nokogiri::XML::Comment.new(frag,'moo')

data/test/xml/test_entity_reference.rb CHANGED

@@ -141,7 +141,7 @@ EOF
         end
         assert_kind_of Nokogiri::XML::EntityReference, doc.xpath('//body').first.children.first
         if Nokogiri.uses_libxml?
-          assert_equal ["Entity 'bar' not defined"], doc.errors.map(&:to_s)
+          assert_equal ["5:14: ERROR: Entity 'bar' not defined"], doc.errors.map(&:to_s)
         end
       end
@@ -159,7 +159,7 @@ EOF
         end
         assert_kind_of Nokogiri::XML::EntityReference, doc.xpath('//body').first.children.first
         if Nokogiri.uses_libxml?
-          assert_equal ["Attempt to load network entity http://foo.bar.com/", "Entity 'bar' not defined"], doc.errors.map(&:to_s)
+          assert_equal ["ERROR: Attempt to load network entity http://foo.bar.com/", "4:34: ERROR: Entity 'bar' not defined"], doc.errors.map(&:to_s)
         else
           assert_equal ["Attempt to load network entity http://foo.bar.com/"], doc.errors.map(&:to_s)
         end

data/test/xml/test_node.rb CHANGED

@@ -91,12 +91,14 @@ module Nokogiri
         assert_equal Nokogiri::XML::ParseOptions::DEFAULT_XML, options.to_i
       end
-      # descriptive, not prescriptive.
-      def test_parse_invalid_html_markup_results_in_empty_nodeset
-        doc = Nokogiri::HTML("<html></html>")
-        nodeset = doc.root.parse "<div><div>a</div><snippet>b</snippet></div>"
-        assert_equal 1, doc.errors.length # "Tag snippet invalid"
+      def test_node_context_parsing_of_malformed_html_fragment
+        doc = HTML.parse "<html><body><div></div></body></html>"
+        context_node = doc.at_css "div"
+        nodeset = context_node.parse("<div </div>")
+        assert_equal 1, doc.errors.length
         assert_equal 1, nodeset.length
+        assert_equal "<div></div>", nodeset.to_s
       end
       def test_node_context_parsing_of_malformed_html_fragment_with_recover_is_corrected
@@ -105,9 +107,10 @@ module Nokogiri
         nodeset = context_node.parse("<div </div>") do |options|
           options.recover
         end
-        assert_equal "<div></div>", nodeset.to_s
         assert_equal 1, doc.errors.length
         assert_equal 1, nodeset.length
+        assert_equal "<div></div>", nodeset.to_s
       end
       def test_node_context_parsing_of_malformed_html_fragment_without_recover_is_not_corrected
@@ -116,6 +119,7 @@ module Nokogiri
         nodeset = context_node.parse("<div </div>") do |options|
           options.strict
         end
         assert_equal 1, doc.errors.length
         assert_equal 0, nodeset.length
       end
@@ -125,7 +129,7 @@ module Nokogiri
         @xml.root.parse('<hello>')
         assert(error_count < @xml.errors.length, "errors should have increased")
       end
       def test_parse_error_on_fragment_with_empty_document
         doc = Document.new
         fragment = DocumentFragment.new(doc, '<foo><bar/></foo>')
@@ -690,14 +694,7 @@ module Nokogiri
       end
       def test_find_by_css_class_with_nonstandard_whitespace
-        doc = Nokogiri::HTML '
-<html>
-  <body>
-    <div class="a
-b"></div>
-  </body>
-</html>
-'
+        doc = Nokogiri::HTML "<div class='a\nb'></div>"
         assert_not_nil doc.at_css(".b")
       end

data/test/xml/test_node_reparenting.rb CHANGED

@@ -418,6 +418,20 @@ module Nokogiri
             assert_equal "text nodefoo <p></p> bar", xml.root.children.to_html
           end
+          it 'should append a text node before an existing non text node' do
+            xml = Nokogiri::XML %Q(<root><p>foo</p><p>bar</p></root>)
+            p = xml.at_css 'p'
+            p.add_next_sibling 'a'
+            assert_equal '<root><p>foo</p>a<p>bar</p></root>', xml.root.to_s
+          end
+          it 'should append a text node before an existing text node' do
+            xml = Nokogiri::XML %Q(<root><p>foo</p>after</root>)
+            p = xml.at_css 'p'
+            p.add_next_sibling 'x'
+            assert_equal '<root><p>foo</p>xafter</root>', xml.root.to_s
+          end
           describe "with a text node after" do
             it "should not defensively dup the 'after' text node" do
               xml = Nokogiri::XML %Q(<root>before<p></p>after</root>)

data/test/xml/test_node_set.rb CHANGED

@@ -252,12 +252,14 @@ module Nokogiri
         assert_equal 2, node_set.first(2).length
       end
-      def test_dup
-        assert node_set = @xml.xpath('//employee')
-        dup = node_set.dup
-        assert_equal node_set.length, dup.length
-        node_set.zip(dup).each do |a,b|
-          assert_equal a, b
+      [:dup, :clone].each do |method_name|
+        define_method "test_#{method_name}" do
+          assert node_set = @xml.xpath('//employee')
+          duplicate = node_set.send(method_name)
+          assert_equal node_set.length, duplicate.length
+          node_set.zip(duplicate).each do |a,b|
+            assert_equal a, b
+          end
         end
       end

data/test/xml/test_reader.rb CHANGED

@@ -185,6 +185,25 @@ module Nokogiri
         assert 1, reader.errors.length
       end
+      def test_errors_is_an_array
+        reader = Nokogiri::XML::Reader(StringIO.new('&bogus;'))
+        assert_raises(SyntaxError) {
+          reader.read
+        }
+        assert_equal [SyntaxError], reader.errors.map(&:class)
+      end
+      def test_pushing_to_non_array_raises_TypeError
+        skip "TODO: JRuby ext does not internally call `errors`" if Nokogiri.jruby?
+        reader = Nokogiri::XML::Reader(StringIO.new('&bogus;'))
+        def reader.errors
+          1
+        end
+        assert_raises(TypeError) {
+          reader.read
+        }
+      end
       def test_attributes?
         reader = Nokogiri::XML::Reader.from_memory(<<-eoxml)
         <x xmlns:tenderlove='http://tenderlovemaking.com/'>

data/test/xml/test_syntax_error.rb CHANGED

@@ -8,23 +8,29 @@ module Nokogiri
         assert_equal 'hello', error.message
       end
-      def test_pushing_to_array
-        reader = Nokogiri::XML::Reader(StringIO.new('&bogus;'))
-        assert_raises(SyntaxError) {
-          reader.read
-        }
-        assert_equal [SyntaxError], reader.errors.map(&:class) unless Nokogiri.jruby? # needs investigation
+      def test_line_column_level_libxml
+        skip unless Nokogiri.uses_libxml?
+        bad_doc = Nokogiri::XML('test')
+        error = bad_doc.errors.first
+        assert_equal "1:1: FATAL: Start tag expected, '<' not found", error.message
+        assert_equal 1, error.line
+        assert_equal 1, error.column
+        assert_equal 3, error.level
       end
-      def test_pushing_to_non_array
-        reader = Nokogiri::XML::Reader(StringIO.new('&bogus;'))
-        def reader.errors
-          1
-        end
-        assert_raises(TypeError) {
-          reader.read
-        }
-      end unless Nokogiri.jruby? # which does not internally call `errors`
+      def test_line_column_level_jruby
+        skip unless Nokogiri.jruby?
+        bad_doc = Nokogiri::XML('<root>test</bar>')
+        error = bad_doc.errors.first
+        assert_equal "The element type \"root\" must be terminated by the matching end-tag \"</root>\".", error.message
+        assert_equal nil, error.line
+        assert_equal nil, error.column
+        assert_equal nil, error.level
+      end
     end
   end
 end

data/test/xml/test_unparented_node.rb CHANGED

@@ -223,12 +223,12 @@ module Nokogiri
         right_space.add_next_sibling(left_space)
         assert_equal left_space, right_space
       end
       def test_add_next_sibling_to_root_raises_exception
         xml = Nokogiri::XML(<<-eoxml)
         <root />
         eoxml
         node = Nokogiri::XML::Node.new 'child', xml
         assert_raise(ArgumentError) do
@@ -240,7 +240,7 @@ module Nokogiri
         xml = Nokogiri::XML(<<-eoxml)
         <root />
         eoxml
         node = Nokogiri::XML::Node.new 'child', xml
         assert_raise(ArgumentError) do
@@ -248,17 +248,60 @@ module Nokogiri
         end
       end
-      def test_add_pi_as_previous_and_next_sibling_to_root_is_ok
+      def test_document_root_can_have_a_comment_sibling_via_add_child
+        doc = Nokogiri::XML "<root>foo</root>"
+        comment = Nokogiri::XML::Comment.new(doc, "this is a comment")
+        doc.add_child comment
+        assert_equal [doc.root, comment], doc.children.to_a
+      end
+      def test_document_root_can_have_a_comment_sibling_via_prepend_child
+        doc = Nokogiri::XML "<root>foo</root>"
+        comment = Nokogiri::XML::Comment.new(doc, "this is a comment")
+        doc.prepend_child comment
+        assert_equal [comment, doc.root], doc.children.to_a
+      end
+      def test_document_root_can_have_a_comment_sibling_via_add_next_sibling
+        doc = Nokogiri::XML "<root>foo</root>"
+        comment = Nokogiri::XML::Comment.new(doc, "this is a comment")
+        doc.root.add_next_sibling comment
+        assert_equal [doc.root, comment], doc.children.to_a
+      end
+      def test_document_root_can_have_a_comment_sibling_via_add_previous_sibling
+        doc = Nokogiri::XML "<root>foo</root>"
+        comment = Nokogiri::XML::Comment.new(doc, "this is a comment")
+        doc.root.add_previous_sibling comment
+        assert_equal [comment, doc.root], doc.children.to_a
+      end
+      def test_document_root_can_have_a_processing_instruction_sibling_via_add_child
         doc = Nokogiri::XML "<root>foo</root>"
         pi = Nokogiri::XML::ProcessingInstruction.new(doc, "xml-stylesheet", %q{type="text/xsl" href="foo.xsl"})
-        doc.root.add_previous_sibling pi
-        expected_doc = %Q{<?xml version="1.0"?>\n<?xml-stylesheet type="text/xsl" href="foo.xsl"?>\n<root>foo</root>}
-        assert_includes doc.to_xml, expected_doc
+        doc.add_child pi
+        assert_equal [doc.root, pi], doc.children.to_a
+      end
+      def test_document_root_can_have_a_processing_instruction_sibling_via_prepend_child
+        doc = Nokogiri::XML "<root>foo</root>"
+        pi = Nokogiri::XML::ProcessingInstruction.new(doc, "xml-stylesheet", %q{type="text/xsl" href="foo.xsl"})
+        doc.prepend_child pi
+        assert_equal [pi, doc.root], doc.children.to_a
+      end
-        pi2 = Nokogiri::XML::ProcessingInstruction.new(doc, "xml-stylesheet", %q{type="text/xsl" href="foo2.xsl"})
-        pi.add_next_sibling pi2
-        expected_doc = %Q{<?xml version="1.0"?>\n<?xml-stylesheet type="text/xsl" href="foo.xsl"?>\n<?xml-stylesheet type="text/xsl" href="foo2.xsl"?>\n<root>foo</root>}
-        assert_includes doc.to_xml, expected_doc
+      def test_document_root_can_have_a_processing_instruction_sibling_via_add_next_sibling
+        doc = Nokogiri::XML "<root>foo</root>"
+        pi = Nokogiri::XML::ProcessingInstruction.new(doc, "xml-stylesheet", %q{type="text/xsl" href="foo.xsl"})
+        doc.root.add_next_sibling pi
+        assert_equal [doc.root, pi], doc.children.to_a
+      end
+      def test_document_root_can_have_a_processing_instruction_sibling_via_add_previous_sibling
+        doc = Nokogiri::XML "<root>foo</root>"
+        pi = Nokogiri::XML::ProcessingInstruction.new(doc, "xml-stylesheet", %q{type="text/xsl" href="foo.xsl"})
+        doc.root.add_previous_sibling pi
+        assert_equal [pi, doc.root], doc.children.to_a
       end
       def test_find_by_css_with_tilde_eql