[Bug] Decoding problem of UTF-8 strings (used e.g. for serialization in session)

[lhaemmerle]

Environment:

• LDAP Server Type: OpenLDAP
• PHP Version: 8.1

Describe the bug: We want to store LDAP user objects (extends LdapRecord\Models\Model) in the Laravel session (serialization). However, this causes the problem that LDAP users with names containing Umlaut characters (e.g. "René") are wrongly decoded when the session is loaded (=LDAP user unserialized) again. The UTF-8 string "René" is considered an 'ISO-8859-1' encoded string and converted to UTF-8, which then becomes "René".

It seems that the cause for this a bug in the function decodeValue(string $value) in HasAttributes.php: https://github.com/DirectoryTree/LdapRecord/blob/master/src/Models/Concerns/HasAttributes.php#L223

I suspect the condition in

if (MbString::isLoaded() && MbString::isUtf8($value)) {
      return mb_convert_encoding($value, 'UTF-8', 'ISO-8859-1');
}

Should rather be (notice the !):

if (MbString::isLoaded() && !MbString::isUtf8($value)) {
      return mb_convert_encoding($value, 'UTF-8', 'ISO-8859-1');
}

because it makes no sense to convert a string that is already utf-8.

[stevebauman]

Ah thanks @lhaemmerle! I'll patch this now and add a test to ensure this works properly.

[stevebauman]

Ok I've just released v3.5.1 with a patch for this issue. I've added test assertions to replicate this bug and ensure it's working properly.

Please run composer update and you're all set! 👍

Appreciate the detailed report 🙏

[lhaemmerle]

Great, thank your for the quick fix!