Turkish character problem in clean url 1.5.3.1

[tebdilikiyafet]

There was no problem in 1.5.1 version in Turkish characters like ''çşığüö'' and they were convert to ''csiguo'' . But now in 1.5.3.1 version it doesn't convert ''ş'' and ''ğ'' to ''s'' and ''g''.

 

Is this convertion done by classes/tool.php ? There is a replacement code like this in the file:

 

public static function replaceAccentedChars($str)
{
 $patterns = array(
  /* Lowercase */
  '/[\x{0105}\x{00E0}\x{00E1}\x{00E2}\x{00E3}\x{00E4}\x{00E5}]/u',
  '/[\x{00E7}\x{010D}\x{0107}]/u',
  '/[\x{010F}]/u',
  '/[\x{00E8}\x{00E9}\x{00EA}\x{00EB}\x{011B}\x{0119}]/u',
  '/[\x{00EC}\x{00ED}\x{00EE}\x{00EF}]/u',
  '/[\x{0142}\x{013E}\x{013A}]/u',
  '/[\x{00F1}\x{0148}]/u',
  '/[\x{00F2}\x{00F3}\x{00F4}\x{00F5}\x{00F6}\x{00F8}]/u',
  '/[\x{0159}\x{0155}]/u',
  '/[\x{015B}\x{0161}]/u',
  '/[\x{00DF}]/u',
  '/[\x{0165}]/u',
  '/[\x{00F9}\x{00FA}\x{00FB}\x{00FC}\x{016F}]/u',
  '/[\x{00FD}\x{00FF}]/u',
  '/[\x{017C}\x{017A}\x{017E}]/u',
  '/[\x{00E6}]/u',
  '/[\x{0153}]/u',
  /* Uppercase */
  '/[\x{0104}\x{00C0}\x{00C1}\x{00C2}\x{00C3}\x{00C4}\x{00C5}]/u',
  '/[\x{00C7}\x{010C}\x{0106}]/u',
  '/[\x{010E}]/u',
  '/[\x{00C8}\x{00C9}\x{00CA}\x{00CB}\x{011A}\x{0118}]/u',
  '/[\x{0141}\x{013D}\x{0139}]/u',
  '/[\x{00D1}\x{0147}]/u',
  '/[\x{00D3}]/u',
  '/[\x{0158}\x{0154}]/u',
  '/[\x{015A}\x{0160}]/u',
  '/[\x{0164}]/u',
  '/[\x{00D9}\x{00DA}\x{00DB}\x{00DC}\x{016E}]/u',
  '/[\x{017B}\x{0179}\x{017D}]/u',
  '/[\x{00C6}]/u',
  '/[\x{0152}]/u');
 $replacements = array(
   'a', 'c', 'd', 'e', 'i', 'l', 'n', 'o', 'r', 's', 'ss', 't', 'u', 'y', 'z', 'ae', 'oe',
   'A', 'C', 'D', 'E', 'L', 'N', 'O', 'R', 'S', 'T', 'U', 'Z', 'AE', 'OE'
  );
 return preg_replace($patterns, $replacements, $str);
}

 

What should I change to make the correct url replacement?

[tebdilikiyafet]

Is there any suggestion for that? Because it was ok in 1.5.1.

 

Now 'ş' and 'ğ' are not converting to 's' and 'g' ? Do I need to add character codes for ş and ğ? How can I solve this?

[benjamin utterback]

Hi, thank you for the message.

 

1.5.4 fixes this issue. However, you should use the newest version 1.5.4.1 as it has critical updates.

 

 

Happy Selling!

[tebdilikiyafet]

Which files do this convertions? classes/tools.php? I have tired 1.5.4.1 but it does not convert them to latin characters. It makes the url with the same chars.

 

For example product name Şerğınçörü and url must be like -> sitename/sergincoru only converted to latin chars.

 

But 1.5.4.1 makes non-latin chars on the url like: sitename/şerğınçörü

 

I don't want non-latin chars in the url. ıt was ok 1.4.x version and also 1.5.1 but it is a bug now.

 

moreover only 2character is the problem, "ş" and "ğ" problem. others are true.

 

Which file should I edit for the true convertion? I think I need to add these two characters' utf codes to tools.php? Or some file else?

Edited April 26, 2013 by tebdilikiyafet (see edit history)

[benjamin utterback]

Hi tebdilikiyatfet, what is the problem with those two characters, they do not work?

[tebdilikiyafet]

There are 6 non-latin character in Turkish these are : ş ğ ç ö ı ü

 

In the friendly urls these characters are converting to latin chars by prestashop -> s g c o i u

 

The problem is that prestashop don't convert "ş" and "ğ" to latin chars in the url. It deletes these two characters.

 

If product name Şerğınçörü it makes the friendly url : sitename/erincoru . As you see it deletes them.

 

This is the problem.

[tebdilikiyafet]

I think the problem is here:

 

// Remove all non-whitelist chars.
 if (Configuration::get('PS_ALLOW_ACCENTED_CHARS_URL'))
  $str = preg_replace('/[^a-zA-Z0-9\s\'\:\/\[\]-\pL]/u', '', $str);
 else
  $str = preg_replace('/[^a-zA-Z0-9\s\'\:\/\[\]-]/','', $str);

 $str = preg_replace('/[\s\'\:\/\[\]-]+/', ' ', $str);
 $str = str_replace(array(' ', '/'), '-', $str);
 // If it was not possible to lowercase the string with mb_strtolower, we do it after the transformations.
 // This way we lose fewer special chars.
 if (!function_exists('mb_strtolower'))
  $str = strtolower($str);
 return $str;
}
/**
 * Replace all accented chars by their equivalent non accented chars.
 *
 * @param string $str
 * @return string
 */
public static function replaceAccentedChars($str)
{
 $patterns = array(
  /* Lowercase */
  '/[\x{0105}\x{00E0}\x{00E1}\x{00E2}\x{00E3}\x{00E4}\x{00E5}]/u',

 

As you see it deletes ''Remove all non-whitelist chars'' It is needed to add these 2 chars to whitelist, but how?

[prestashopninja]

If you don't want to upgrade and just save the day:

1. download the attached file and upload it on your PS installation, under overrides/classes directory.

2. delete cache/class_index.php file (don't be afraid, it will be regenerated)

 

This is the same method as the original, with additions for turkish special characters.

 

I hope this helps.

Tools.php.zip

Edited April 27, 2013 by prestashopninja (see edit history)