Dave Page wrote:
>It seems we all seem to think that client_encoding should be set to
>UNICODE in all cases for unicode builds and SQL_ASCII for others,
>therefore I've backed out the previous changes and implemented this
>behaviour instead.
>
Looks good to me.
I've changed libpq handling so utf-8 is correctly converted to utf-16. I
believe encoding is now handled correctly for unicode builds. A current
win32 build is uploaded to snake.pgadmin.org.
Client encoding setting might still be useful for non-unicode builds, if
we want to support them at all. The list of selectable client encodings
would have to be restricted to 8-bit/char codes.
Actually, I'd prefer to recommend unicode-builds as *the* one and only
solution for multi-encoding environments, leaving the non-unicode build
pinned to sql_ascii.
Regards,
Andreas