A Russian user kindly alerted us to a bug that we have now fixed: filenames that used Cyrillic characters were not getting uploaded correctly, because of an encoding problem.

The problem had to do with Kerika not using a sufficiently large character set, which resulted in some languages not being supported properly.

We have fixed that by moving away from the UTF8 character encoding to UTF8MB4, which is an encoding format used by MySQL databases.

UTF8 uses a maximum of 3 bytes per character, which limits the range of languages that can be supported.  UTF8MB4, however, uses a maximum of 4 bytes per character which considerably extends the range of possible characters.