Tiago Gomes wrote:

Can i omit non-unicode characters to avoid the error: Server Error! (OK). The sent-mail folder from another webmail are with non-unicode characters. How can omit these characters?

Supplement: this happens in IE8 when rcmail_js_message_list() returns data with non-unicode characters.

I got function to remove those characters. It fixes the issue, but I'm not sure if this is one and only solution.

function rc_utf8_clean($str) { // iconv is 10x faster if (function_exists('iconv')) return iconv('UTF8', 'UTF8//IGNORE', $str);

     $regexp = '/^('.

// '[\x00-\x7F]'. // UTF8-1 '|[\xC2-\xDF][\x80-\xBF]'. // UTF8-2 '|\xE0[\xA0-\xBF][\x80-\xBF]'. // UTF8-3 '|[\xE1-\xEC][\x80-\xBF][\x80-\xBF]'. // UTF8-3 '|\xED[\x80-\x9F][\x80-\xBF]'. // UTF8-3 '|[\xEE-\xEF][\x80-\xBF][\x80-\xBF]'. // UTF8-3 '|\xF0[\x90-\xBF][\x80-\xBF][\x80-\xBF]'. // UTF8-4 '|[\xF1-\xF3][\x80-\xBF][\x80-\xBF][\x80-\xBF]'.// UTF8-4 '|\xF4[\x80-\x8F][\x80-\xBF][\x80-\xBF]'. // UTF8-4 ')$/'; $seq = ''; $out = '';

     for ($i = 0, $len = strlen($str)-1; $i < $len; $i++) {
         $chr = $str[$i];
         $ord = ord($chr);
         // 1-byte character
         if ($ord <= 0x7F) {
             if ($seq)
                 $out .= preg_match($regexp, $seq) ? $seq : '';
             $seq = '';
             $out .= $chr;
         // first (or second) byte of multibyte sequence
         } else if ($ord >= 0xC0) {
             if (strlen($seq)>1) {
                     $out .= preg_match($regexp, $seq) ? $seq : '';
                     $seq = '';
             } else if ($seq && ord($seq) < 0xC0) {
                     $seq = '';
             }
             $seq .= $chr;
         // next byte of multibyte sequence
         } else if ($seq) {
             $seq .= $chr;
         }
     }

     if ($seq)
             $out .= preg_match($regexp, $seq) ? $seq : '';

     return $out;

}

Thomas Bruederli wrote:

Hi, can you send us a sample message that causes theses problems? I'd like to reproduce and test it in order to find a solution. The rc_utf8_clean() funtion will probably solve the problem but it also could have side-effects.

I've got one. Attached.