Fix comments search result to work with multibyte strings

Currently, the searching in comments breaks up, if comments contain multibyte characters and string manipulation logic in getRelevantMessagePart happens to cut through them rendering the resulting string invalid (not UTF-8 compliant). This patch replaces all string manipulating functions in this code to multibyte aware ones.

Signed-off-by: Michał Węgrzynek <michal.wegrzynek@malloc.com.pl>