Tatsuo Ishii <ishii@postgresql.org> writes:
> BTW, do we have to modify pg_dump or pg_restore so that it can
> automatically adjust JOHAB to UTF8 (it's the only safe encoding
> compatible with JOHAB)? I'm not sure it's worth the trouble. Maybe
> documenting in the release note is enough?
Do we actually need to do anything? Dumps taken in client_encoding
JOHAB could exist regardless of the source server_encoding --- the
same is true of other client-only encodings. Such dumps should load
fine into a UTF8 server_encoding database, as long as we have the right
conversion available.
I can imagine someone wanting to take a dump in a client-only encoding
for other reasons (export of the data to somewhere else, say) so I don't
think pg_dump should try to prevent it.
regards, tom lane