From d19c0f9f309cbe63411a8ddcbbda3daf7461a30d Mon Sep 17 00:00:00 2001 From: Aleksander Machniak Date: Thu, 12 Dec 2013 08:58:54 +0100 Subject: In normalize_string() replace 4-byte unicode characters with '?' character. These are not supported in default utf-8 charset on mysql, the chance we'd need them in searching is very low. --- program/lib/Roundcube/rcube_utils.php | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'program/lib/Roundcube/rcube_utils.php') diff --git a/program/lib/Roundcube/rcube_utils.php b/program/lib/Roundcube/rcube_utils.php index 27a618d83..db41a6e86 100644 --- a/program/lib/Roundcube/rcube_utils.php +++ b/program/lib/Roundcube/rcube_utils.php @@ -912,10 +912,20 @@ class rcube_utils * * @param string Input string (UTF-8) * @param boolean True to return list of words as array + * * @return mixed Normalized string or a list of normalized tokens */ public static function normalize_string($str, $as_array = false) { + // replace 4-byte unicode characters with '?' character, + // these are not supported in default utf-8 charset on mysql, + // the chance we'd need them in searching is very low + $str = preg_replace('/(' + . '\xF0[\x90-\xBF][\x80-\xBF]{2}' + . '|[\xF1-\xF3][\x80-\xBF]{3}' + . '|\xF4[\x80-\x8F][\x80-\xBF]{2}' + . ')/', '?', $str); + // split by words $arr = self::tokenize_string($str); -- cgit v1.2.3