Reworked token generation for elements.

This emits separate tokens for the start tag (T_ELEMENT_OPEN) and name
(T_ELEMENT_NAME). This makes it easier to include the namespace of an element
(T_ELEMENT_NS) in the output.

lib/oga/lexer.rl

@@ -230,19 +230,35 @@ module Oga
       #
       # http://www.w3.org/TR/html-markup/syntax.html#syntax-elements
       #
-      element_name  = [a-zA-Z0-9\-_]+;
-      element_start = '<' element_name;
 
-      # First emit the token, then advance the column. This way the column
-      # number points to the < and not the "p" in <p>.
+      # Action that creates the tokens for the opening tag, name and namespace
+      # (if any). Remaining work is delegated to a dedicated machine.
       action open_element {
-        t(:T_ELEM_OPEN, @ts + 1)
-
+        add_token(:T_ELEM_OPEN, nil)
         advance_column
 
+        # Add the element name. If the name includes a namespace we'll break
+        # the name up into two separate tokens.
+        name = text(@ts + 1)
+
+        if name.include?(':')
+          ns, name = name.split(':')
+
+          add_token(:T_ELEM_NS, ns)
+
+          # Advance the column for the colon (:) that separates the namespace
+          # and element name.
+          advance_column
+        end
+
+        add_token(:T_ELEM_NAME, name)
+
         fcall element;
       }
 
+      element_name  = [a-zA-Z0-9\-_:]+;
+      element_start = '<' element_name;
+
       element_text := |*
         ^'<' => buffer_text;
 
@@ -275,12 +291,13 @@ module Oga
         # Non self-closing elements.
         '</' element_name {
           emit_text_buffer
-          t(:T_ELEM_CLOSE, p)
+          add_token(:T_ELEM_CLOSE, nil)
 
-          # Advance by two to take the closing </ into account. This is done
-          # after emitting tokens to ensure that they point to the start of
-          # the tag.
+          # Advance the column for the </
           advance_column(2)
+
+          # Advance the column for the closing name.
+          advance_column(@te - p)
           fret;
         };
 
@@ -311,9 +328,6 @@ module Oga
         };
 
         element_start => open_element;
-
-        #dquote => { t(:T_DQUOTE) };
-        #squote => { t(:T_SQUOTE) };
       *|;
     }%%
   end # Lexer

spec/oga/lexer/elements_spec.rb

@@ -4,32 +4,36 @@ describe Oga::Lexer do
   context 'elements' do
     example 'lex an opening element' do
       lex('<p>').should == [
-        [:T_ELEM_OPEN, 'p', 1, 1]
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NAME, 'p', 1, 2]
       ]
     end
 
     example 'lex an opening an closing element' do
       lex('<p></p>').should == [
-        [:T_ELEM_OPEN, 'p', 1, 1],
-        [:T_ELEM_CLOSE, 'p', 1, 4]
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NAME, 'p', 1, 2],
+        [:T_ELEM_CLOSE, nil, 1, 4]
       ]
     end
 
     example 'lex a paragraph element with text inside it' do
       lex('<p>Hello</p>').should == [
-        [:T_ELEM_OPEN, 'p', 1, 1],
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NAME, 'p', 1, 2],
         [:T_TEXT, 'Hello', 1, 4],
-        [:T_ELEM_CLOSE, 'p', 1, 9]
+        [:T_ELEM_CLOSE, nil, 1, 9]
       ]
     end
 
     example 'lex a paragraph element with attributes' do
       lex('<p class="foo">Hello</p>').should == [
-        [:T_ELEM_OPEN, 'p', 1, 1],
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NAME, 'p', 1, 2],
         [:T_ATTR, 'class', 1, 4],
         [:T_STRING, 'foo', 1, 10],
         [:T_TEXT, 'Hello', 1, 16],
-        [:T_ELEM_CLOSE, 'p', 1, 21]
+        [:T_ELEM_CLOSE, nil, 1, 21]
       ]
     end
   end
@@ -37,22 +41,26 @@ describe Oga::Lexer do
   context 'nested elements' do
     example 'lex a nested element' do
       lex('<p><a></a></p>').should == [
-        [:T_ELEM_OPEN, 'p', 1, 1],
-        [:T_ELEM_OPEN, 'a', 1, 4],
-        [:T_ELEM_CLOSE, 'a', 1, 7],
-        [:T_ELEM_CLOSE, 'p', 1, 11]
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NAME, 'p', 1, 2],
+        [:T_ELEM_OPEN, nil, 1, 4],
+        [:T_ELEM_NAME, 'a', 1, 5],
+        [:T_ELEM_CLOSE, nil, 1, 7],
+        [:T_ELEM_CLOSE, nil, 1, 11]
       ]
     end
 
     example 'lex nested elements and text nodes' do
       lex('<p>Foo<a>bar</a>baz</p>').should == [
-        [:T_ELEM_OPEN, 'p', 1, 1],
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NAME, 'p', 1, 2],
         [:T_TEXT, 'Foo', 1, 4],
-        [:T_ELEM_OPEN, 'a', 1, 7],
+        [:T_ELEM_OPEN, nil, 1, 7],
+        [:T_ELEM_NAME, 'a', 1, 8],
         [:T_TEXT, 'bar', 1, 10],
-        [:T_ELEM_CLOSE, 'a', 1, 13],
+        [:T_ELEM_CLOSE, nil, 1, 13],
         [:T_TEXT, 'baz', 1, 17],
-        [:T_ELEM_CLOSE, 'p', 1, 20]
+        [:T_ELEM_CLOSE, nil, 1, 20]
       ]
     end
   end
@@ -60,18 +68,31 @@ describe Oga::Lexer do
   context 'void elements' do
     example 'lex a void element' do
       lex('<br />').should == [
-        [:T_ELEM_OPEN, 'br', 1, 1],
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NAME, 'br', 1, 2],
         [:T_ELEM_CLOSE, nil, 1, 6]
       ]
     end
 
     example 'lex a void element with an attribute' do
       lex('<br class="foo" />').should == [
-        [:T_ELEM_OPEN, 'br', 1, 1],
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NAME, 'br', 1, 2],
         [:T_ATTR, 'class', 1, 5],
         [:T_STRING, 'foo', 1, 11],
         [:T_ELEM_CLOSE, nil, 1, 18]
       ]
     end
   end
+
+  context 'elements with namespaces' do
+    example 'lex an element with namespaces' do
+      lex('<foo:p></p>').should == [
+        [:T_ELEM_OPEN, nil, 1, 1],
+        [:T_ELEM_NS, 'foo', 1, 2],
+        [:T_ELEM_NAME, 'p', 1, 6],
+        [:T_ELEM_CLOSE, nil, 1, 8]
+      ]
+    end
+  end
 end