Description:
------------
PHP-version: 5.6.29-0+deb8u1 (on Linux plesk4.duocast.net 3.16.0-4-amd64 #1 SMP Debian 3.16.36-1+deb8u2 (2016-10-19) x86_64)

Some entries obtained from get_html_translation_table have a missing closing semicolon.
E.g. [8402] => &nvlt  (should be <⃒)
[8421] => &bne
etc.

If you run the test script below (not the most sophisticated, but a quicky), the invalid entries will be in red.

Test script:
---------------
$vec = get_html_translation_table( HTML_ENTITIES, ENT_QUOTES | ENT_HTML5);
foreach ($vec as $key => $value) {
   $dec = Utf2code( $key );
   if (substr($value,-1,1)!=';')  echo "<br>[$dec] => &amp;<span class=red>".substr($value,1)."</span> = $value;";
   else echo "<br>[$dec] => &amp;".substr($value,1)." = $value";
}

  function Utf2code( $latin ) {
    $magic = 128;
    $i = $cnt = 0;
    while ($i < strlen($latin) ) {
  		$kar = substr( $latin, $i, 1 );
      $num = ord( $kar );
      if (strlen($latin)==1) return( $num );
      if ($num >= $magic) {
        $cnt++;
        $mask = $magic;  $len = $utf = 0;
        while ($mask & $num)  { $len++;  $mask = $mask/2; };
  			$utf = ($mask-1) & $num;
  			$j = 1;
  			while ($j < $len) {
          $kar = substr( $latin, $i+$j, 1 );
          if ((192 & ord( $kar )) != 128)  break;
  			  $utf = ($utf * 64) + (63 & ord($kar));
  			  $j++;
     		};
     		if (($j==$len) and ($len<=6)) {		//Assume UTF
          return( $utf );
        };
      };
      $i++;
    };
    return( -1 );
  };


Expected result:
----------------
A dump of the html (entity) translation table, with incomplete entries flagged in red.