summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorAleksander Machniak <alec@alec.pl>2013-12-12 08:58:54 +0100
committerAleksander Machniak <alec@alec.pl>2013-12-12 08:58:54 +0100
commitd19c0f9f309cbe63411a8ddcbbda3daf7461a30d (patch)
tree4b5a72bb04d3cb16296d7ab8c5a3218fddaaf9e6
parent7eecf873da8d2f28e20dc8fd0e949e6abc5762b4 (diff)
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.
-rw-r--r--program/lib/Roundcube/rcube_utils.php10
1 files changed, 10 insertions, 0 deletions
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);