Description:
------------
A piece of code that was working perfectly correctly with PHP version 4.2.3 is now displaying erroneous characters for PHP 4.3.2.

The script in question is a translation form, displaying two languages which may use different charsets.
The code determines the appropriate charset to use for the html header from the two languages. If the charsets are the same, no problem. If they are different, it tests whther the mbstring module is enabled, in which case it uses 'UTF-8' for the html header; otherwise it uses the charset for the second language.

If mbstring was enabled, then the code uses the mb_convert_encoding function to convert the text strings for display to UTF-8...


Reproduce code:
---------------
$basecharset =  'ISO-8859-1';
$charset = 'ISO-8859-2';

$convertcharsets = ($basecharset != $charset);
if ($convertcharsets) {
   if (function_exists('mb_convert_encoding')) {
      formheader('UTF-8');
   } else {
      $convertcharsets = false;
      formheader($charset);
   }
} else {
   formheader($basecharset);
}

echo charsetText('Fran?ais',$convertcharset,$basecharset)
echo '<br />';
echo charsetText('Polska',$convertcharset,$charset)
echo '<br />';

function charsetText($text,$convertcharset,$fromcharset)
{
   $returntext = $text;
   if ($convertcharset) { $returntext = mb_convert_encoding($returntext,"UTF-8",$fromcharset); }
   return $returntext;
} // function charsetText()


Expected result:
----------------
Fran?ais
Polska



Actual result:
--------------
Fran?&sect;ais
Polska

The first odd character in FranXXais is A with a tilde; the second is the HTML &sect; character