Support whitespace in element closing tags

Fixes #108

ext/ragel/base_lexer.rl

@@ -368,10 +368,12 @@
         fnext element_name;
     }
 
-    action close_element {
+    action start_close_element {
-        callback(id_on_element_end, data, encoding, mark, te - 1);
+        fnext element_close;
+    }
 
-        mark = 0;
+    action close_element {
+        callback(id_on_element_end, data, encoding, ts, te);
     }
 
     action close_element_fnext_main {
@@ -381,10 +383,7 @@
     }
 
     element_start = '<' ident_char;
-
+    element_end   = '</';
-    element_end = '</' %{ mark = p; } identifier '>'
-                | '</' identifier ':' %{ mark = p; } identifier '>'
-                ;
 
     # Machine used for lexing the name/namespace of an element.
     element_name := |*
@@ -398,6 +397,28 @@
         };
     *|;
 
+    # Machine used for lexing the closing tag of an element
+    element_close := |*
+        # namespace prefixes, currently not used but allows the rule below it
+        # to be used for the actual element name.
+        identifier ':';
+
+        identifier => close_element;
+
+        '>' => {
+            if ( lines > 0 )
+            {
+                advance_line(lines);
+
+                lines = 0;
+            }
+
+            fnext main;
+        };
+
+        any $count_newlines;
+    *|;
+
     # Characters that can be used for unquoted HTML attribute values.
     # See https://html.spec.whatwg.org/multipage/introduction.html#intro-early-example
     # for more info.
@@ -587,7 +608,7 @@
         cdata_start    => start_cdata;
         proc_ins_start => start_proc_ins;
         element_start  => start_element;
-        element_end    => close_element;
+        element_end    => start_close_element;
         any            => start_text;
     *|;
 }%%

spec/oga/xml/lexer/elements_spec.rb

@@ -73,6 +73,30 @@ describe Oga::XML::Lexer do
         [:T_ELEM_END, nil, 2]
       ]
     end
+
+    it 'lexes an element with a space in the closing tag' do
+      lex("<foo></foo >bar").should == [
+        [:T_ELEM_NAME, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_TEXT, 'bar', 1]
+      ]
+    end
+
+    it 'lexes an element with a newline in the closing tag' do
+      lex("<foo></foo\n>bar").should == [
+        [:T_ELEM_NAME, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_TEXT, 'bar', 2]
+      ]
+    end
+
+    it 'lexes an element with a newline in the closing tag using an IO as input' do
+      lex(StringIO.new("<foo></foo\n>bar")).should == [
+        [:T_ELEM_NAME, 'foo', 1],
+        [:T_ELEM_END, nil, 1],
+        [:T_TEXT, 'bar', 2]
+      ]
+    end
   end
 
   describe 'elements with attributes' do

spec/oga/xml/lexer/general_spec.rb

@@ -30,10 +30,6 @@ describe Oga::XML::Lexer do
       lex('>').should == [[:T_TEXT, '>', 1]]
     end
 
-    it 'lexes </ as regular text' do
-      lex('</').should == [[:T_TEXT, '</', 1]]
-    end
-
     it 'lexes <! as regular text' do
       lex('<!').should == [[:T_TEXT, '<!', 1]]
     end