self-close certain XML/HTML elements.

When an XML element has no child nodes a self-closing tag is used. When parsing
documents/elements in HTML mode this is only done if the element is a so called
"void element" (e.g. <link> tags).

This fixes #46.

lib/oga.rb

@@ -18,6 +18,7 @@ if RUBY_PLATFORM == 'java'
 end
 #:nocov:
 
+require_relative 'oga/xml/html_void_elements'
 require_relative 'oga/xml/querying'
 require_relative 'oga/xml/traversal'
 require_relative 'oga/xml/node'

lib/oga/xml/element.rb

@@ -211,7 +211,12 @@ module Oga
       # @return [String]
       #
       def to_xml
-        ns    = namespace_name ? "#{namespace_name}:" : ''
+        if namespace_name
+          full_name = "#{namespace_name}:#{name}"
+        else
+          full_name = name
+        end
+
         body  = children.map(&:to_xml).join('')
         attrs = ''
 
@@ -219,7 +224,11 @@ module Oga
           attrs << " #{attr.to_xml}"
         end
 
-        return "<#{ns}#{name}#{attrs}>#{body}</#{ns}#{name}>"
+        if self_closing?
+          return "<#{full_name}#{attrs} />"
+        else
+          return "<#{full_name}#{attrs}>#{body}</#{full_name}>"
+        end
       end
 
       ##
@@ -278,6 +287,23 @@ module Oga
         return merged
       end
 
+      ##
+      # Returns `true` if the element is a self-closing element.
+      #
+      # @return [TrueClass|FalseClass]
+      #
+      def self_closing?
+        self_closing = children.empty?
+        root         = root_node
+
+        if root.is_a?(Document) and root.type == :html \
+        and !HTML_VOID_ELEMENTS.include?(name)
+          self_closing = false
+        end
+
+        return self_closing
+      end
+
       private
 
       ##

lib/oga/xml/html_void_elements.rb

@@ -0,0 +1,28 @@
+module Oga
+  module XML
+    ##
+    # Names of the HTML void elements that should be handled when HTML lexing
+    # is enabled.
+    #
+    # @return [Set]
+    #
+    HTML_VOID_ELEMENTS = Set.new([
+      'area',
+      'base',
+      'br',
+      'col',
+      'command',
+      'embed',
+      'hr',
+      'img',
+      'input',
+      'keygen',
+      'link',
+      'meta',
+      'param',
+      'source',
+      'track',
+      'wbr'
+    ])
+  end # XML
+end # Oga

lib/oga/xml/lexer.rb

@@ -40,31 +40,6 @@ module Oga
     class Lexer
       attr_reader :html
 
-      ##
-      # Names of the HTML void elements that should be handled when HTML lexing
-      # is enabled.
-      #
-      # @return [Set]
-      #
-      HTML_VOID_ELEMENTS = Set.new([
-        'area',
-        'base',
-        'br',
-        'col',
-        'command',
-        'embed',
-        'hr',
-        'img',
-        'input',
-        'keygen',
-        'link',
-        'meta',
-        'param',
-        'source',
-        'track',
-        'wbr'
-      ])
-
       ##
       # @param [String|IO] data The data to lex. This can either be a String or
       #  an IO instance.

spec/oga/xml/element_spec.rb

@@ -267,17 +267,18 @@ describe Oga::XML::Element do
 
   context '#to_xml' do
     example 'generate the corresponding XML' do
-      described_class.new(:name => 'p').to_xml.should == '<p></p>'
+      described_class.new(:name => 'p').to_xml.should == '<p />'
     end
 
     example 'include the namespace if present' do
       instance = described_class.new(
         :name           => 'p',
         :namespace_name => 'foo',
-        :namespaces     => {'foo' => Oga::XML::Namespace.new(:name => 'foo')}
+        :namespaces     => {'foo' => Oga::XML::Namespace.new(:name => 'foo')},
+        :children       => [Oga::XML::Text.new(:text => 'Foo')]
       )
 
-      instance.to_xml.should == '<foo:p></foo:p>'
+      instance.to_xml.should == '<foo:p>Foo</foo:p>'
     end
 
     example 'include a single attribute if present' do
@@ -288,7 +289,7 @@ describe Oga::XML::Element do
         ]
       )
 
-      instance.to_xml.should == '<p key="value"></p>'
+      instance.to_xml.should == '<p key="value" />'
     end
 
     example 'include multiple attributes if present' do
@@ -300,7 +301,7 @@ describe Oga::XML::Element do
         ]
       )
 
-      instance.to_xml.should == '<p key1="value1" key2="value2"></p>'
+      instance.to_xml.should == '<p key1="value1" key2="value2" />'
     end
 
     example 'include the child nodes if present' do
@@ -319,7 +320,21 @@ describe Oga::XML::Element do
         :namespaces => {'xmlns' => namespace}
       )
 
-      instance.to_xml.should == '<foo></foo>'
+      instance.to_xml.should == '<foo />'
+    end
+
+    example 'generate the XML for the HTML <script> element' do
+      element  = described_class.new(:name => 'script')
+      document = Oga::XML::Document.new(:type => :html, :children => [element])
+
+      element.to_xml.should == '<script></script>'
+    end
+
+    example 'generate the XML for the HTML <link> element' do
+      element  = described_class.new(:name => 'link')
+      document = Oga::XML::Document.new(:type => :html, :children => [element])
+
+      element.to_xml.should == '<link />'
     end
   end
 
@@ -411,4 +426,32 @@ describe Oga::XML::Element do
       @parent_ns['baz'].uri.should == 'yyy'
     end
   end
+
+  context '#self_closing?' do
+    example 'return true for an empty XML element' do
+      described_class.new(:name => 'foo').should be_self_closing
+    end
+
+    example 'return false for a non empty XML element' do
+      text = Oga::XML::Text.new(:text => 'bar')
+      node = described_class.new(:name => 'foo', :children => [text])
+
+      node.should_not be_self_closing
+    end
+
+    example 'return true for an HTML void element' do
+      element  = described_class.new(:name => 'link')
+      document = Oga::XML::Document.new(:type => :html, :children => [element])
+
+      element.should be_self_closing
+    end
+
+    example 'return false for a non empty HTML element' do
+      text     = Oga::XML::Text.new(:text => 'alert()')
+      element  = described_class.new(:name => 'script', :children => [text])
+      document = Oga::XML::Document.new(:type => :html, :children => [element])
+
+      element.should_not be_self_closing
+    end
+  end
 end