Word breaking works wrong with multibyte characters

When non-latin characters (e. g. cyrillic) are used, the wbr tag appears more often than it is needed and breaks UTF-8 characters. The possible solution is to change function _mobileplugin_break_long_words() to

function _mobileplugin_break_long_words($str, $max) {
	$res = mb_ereg_replace("\w{".$max."}", "\\0"."<wbr />", $str);
	return $res;
}