Merge pull request #218 from ohader/comment-handling

Address comment handling issues

Author: goetas

src/HTML5/Parser/Scanner.php

@@ -104,7 +104,7 @@ public function position()
      */
     public function peek()
     {
-        if (($this->char + 1) <= $this->EOF) {
+        if (($this->char + 1) < $this->EOF) {
             return $this->data[$this->char + 1];
         }
 

src/HTML5/Parser/Tokenizer.php

@@ -712,18 +712,24 @@ protected function isCommentEnd()
             return true;
         }
 
-        // If it doesn't start with -, not the end.
-        if ('-' != $tok) {
+        // If next two tokens are not '--', not the end.
+        if ('-' != $tok || '-' != $this->scanner->peek()) {
             return false;
         }
 
-        // Advance one, and test for '->'
-        if ('-' == $this->scanner->next() && '>' == $this->scanner->peek()) {
+        $this->scanner->consume(2); // Consume '-' and one of '!' or '>'
+
+        // Test for '>'
+        if ('>' == $this->scanner->current()) {
+            return true;
+        }
+        // Test for '!>'
+        if ('!' == $this->scanner->current() && '>' == $this->scanner->peek()) {
             $this->scanner->consume(); // Consume the last '>'
             return true;
         }
-        // Unread '-';
-        $this->scanner->unconsume(1);
+        // Unread '-' and one of '!' or '>';
+        $this->scanner->unconsume(2);
 
         return false;
     }

test/HTML5/Parser/TokenizerTest.php

@@ -197,11 +197,17 @@ public function testComment()
     {
         $good = array(
             '<!--easy-->' => 'easy',
+            '<!--easy--!>' => 'easy',
             '<!-- 1 > 0 -->' => ' 1 > 0 ',
+            '<!-- 1 > 0 --!>' => ' 1 > 0 ',
             '<!-- --$i -->' => ' --$i ',
+            '<!-- --$i --!>' => ' --$i ',
             '<!----$i-->' => '--$i',
+            '<!----$i--!>' => '--$i',
             "<!--\nHello World.\na-->" => "\nHello World.\na",
+            "<!--\nHello World.\na--!>" => "\nHello World.\na",
             '<!-- <!-- -->' => ' <!-- ',
+            '<!-- <!-- --!>' => ' <!-- ',
         );
         foreach ($good as $test => $expected) {
             $events = $this->parse($test);
@@ -213,6 +219,7 @@ public function testComment()
             '<!--Hello' => 'Hello',
             "<!--\0Hello" => UTF8Utils::FFFD . 'Hello',
             '<!--' => '',
+            '<!--<!--' => '<!--',
         );
         foreach ($fail as $test => $expected) {
             $events = $this->parse($test);