Hi,
I'm going to add a new function "pg_client_encoding" returning the
current client side encoding name. I know there is a similar
functionality already there in PostgreSQL (show client_encoding) but
it's pain to handle notice message by a program.
Also note that JDBC driver and maybe some other APIs use
getdatabaseencoding, but I think it's not adequate for FE APIs to know
actual encoding passed to FE side, since an encoding conversion might
be made in BE side. For example, if PGCLIENTENCODING is set to SJIS
before starting postmaster, the actual encoding passed to FE would be
SJIS even the database encoding is EUC_JP.
Comments?
--
Tatsuo Ishii