Bug Report

min_length & max_length utf-8 problem
Date:	06/29/2008	Severity:	Minor
Status:	Resolved	Reporter:	Zver1992
Version:	1.6.3
Keywords:	Libraries, Validation Class

Description

Standart PHP function “strlen()“ count length of some utf-8 strings (russian language and some languages cyryllic too) wrong. This function count 1 russian symbol = 2. In manual ( http://www.php.net/strlen ) im found correctly function. Examples:

<?php
function strlen_utf8($str)
{
$i = 0;
$count = 0;
$len = strlen($str);
while($i < $len) {
$chr = ord($str[$i]);
$count++;
$i++;
if($i >= $len) {
break;
}
if($chr & 0x80) {
$chr <<= 1;
while($chr & 0x80) {
$i++;
$chr <<= 1;
}
}
}
return $count;
}
$string = ‘авбгд‘;
echo strlen($string); //10, wrong
echo strlen_utf8($string); //5, right
?>

Im replace all strlen’s for strlen_utf8 in my Validation Class and my script work correctly now

Code Sample

<?php
    function min_length($str, $val)
    {
        if (preg_match("/[^0-9]/", $val))
        {
            return FALSE;
        }
    
        return (strlen($str) < $val) ? FALSE : TRUE;
    }
    $check = min_length('авбгд', 6);
?> 

Expected Result

FALSE

Actual Result

TRUE

Comment on Bug Report

Page 1 of 1 pages

Posted by: Sam Dark on 30 June 2008 10:29am
	It’s not only with strlen. As I know, there are also problems with: trim, ltrim, rtrim str_ireplace ord str_pad str_split strcasecmp strcspn stristr strlen strpos strrev strrpos strspn strtolower, strtoupper substr substr_replace ucfirst, ucwords
Posted by: inparo on 1 July 2008 8:46am
	That is what the multibite string functions are for. So in your case it would be a simple mb_strlen($str, “UTF-8”). Remember that you can add your own validation rules by extending the class, so there’s no need to alter core files.
Posted by: Sam Dark on 1 July 2008 9:05am
	inparo, multibite string functions aren’t perfect either.
Posted by: inparo on 1 July 2008 9:55am
	No, but they solve the ‘bug’ outlined above.
Posted by: Mark (Germany) on 10 July 2008 5:28am
	Woudn’t it be way nicer to let CodeI. handle it? So you can just use the strlen method and it handles itself, which function is the right one? this can be done - like in other cms or cakephp - by simply setting the UTF8 parameter in config - and then checking if the mb_ functions are availible and using them instead of the primary ones. thats, what you usually would expect using this functions.
Posted by: Sam Dark on 10 July 2008 5:34am
	This can be really useful.