"Rolandas Rudomanskis" <rolandasr(at)gmail(dot)com> writes:
> Description: Lithuanian characters doesn't work in queries with
> regular expressions
> The database is set to unicode. Debian locale is utf-8.
Yeah, the regex code doesn't handle localization issues (such as
case-insensitive matching) correctly for multibyte characters.
Can you use a single-byte encoding? Alternatively, I believe
case-insensitive LIKE matching works for this.
regards, tom lane