Re: LATIN2->UTF8 conversation with dblink

Tom Lane wrote:
> Ruzsinszky Attila <ruzsinszky.attila@gmail.com> writes:
>> The DB is the same except the character coding. Source is LATIN2 and
>> the target DB is UTF8.
>> We wrote a trigger to copy the data from source to target with dblink.
>> The problem is the
>> different DB character coding! PGSQL complains about wrong byte order.
>
> Hmm.  You can presumably fix this by setting client_encoding in the
> dblink connection to match the encoding in use in the database it's
> called in.  But I wonder why dblink doesn't just do that for you
> automatically.

Mainly because I never thought about it myself before, and this is the
first time I've seen someone complain ;-)

But if you think automatically setting client encoding is appropriate, I
will make the change. Would it be classified as a bug (and therefore
something to backpatch) or a new feature?

Joe