Inconsistent behavior of regex for UTF-8 input

Currently Regex matching is not consistent between input strings and Zabbix components.
For the test we can use the same pattern "(.)(.)(.)(.)" (without quotes).

1. This pattern works with "abcd" on both Server and Frontend:
   
   
2. This pattern works with "абвг" on Frontend (well, it still splits UTF sequences into separate bytes, but it doesn't produce the error), but fails on Server:
   
   

Currently, in order for regex to work correctly on Server (as well as on Agent and Proxy), utf flag should be set like so "(*UTF)(.)(.)(.)(.)" (without quotes):

It is expected for Server to work with UTF-8 by default.