If utf-8 character is used in all translations, it simplifies coding and fixes some issues.
1. Users can use any character instead of limited set of characters supported by some iso-8859-x charset. Compose conversions always work.
2. User preferences are not corrupted when users change interface language.
3. No Catch22 in database preferences and address books.
(#1520866)
4. Simplified 8bit data checks. After decoding data can contain only utf-8 and html entities.
Forced use of utf-8 will create issues with incorrectly formated MIME messages and compatibility issues with older user preferences. In order to improve CJK support interface must bundle extra decoding functions.
