Encoding conversions in psql

Howdy,

Can anyone explain to me when psql tries to convert between encodings?
It seems to disregard encodings set with SET CLIENT_ENCODING.

The following reproduces the behaviour I'm seeing:

1. create an UNICODE database

2. run the following:
set client_encoding to latin1;
create table bla(a text);
insert into bla values('meëep');

3. try the following from psql:
Welcome to psql 7.3.4, the PostgreSQL interactive terminal.

Type: \copyright for distribution terms
\h for help with SQL commands
\? for help on internal slash commands
\g or terminate with semicolon to execute query
\q to quit

mathijs=# select * from bla;
a
-------
meëep
(1 row)

mathijs=# set client_encoding = latin1;
SET
mathijs=# select * from bla;
a
------
meep
(1 row)

mathijs=# \encoding latin1
mathijs=# select * from bla;
a
-------
meëep
(1 row)

After setting CLIENT_ENCODING, the middle character gets dropped. To me
it seems like psql is considering the data it gets from the server as
UTF8, tries to interpret it as UTF8, sees the ë (which is indeed an
invalid UTF8 character) and drops it.

My question is: why does psql seem to think it's receiving UTF8 data
-after- I've changed the client_encoding. I've checked with a network
sniffer that results returned with or without using \encoding (as
expected) are the same. Is this behaviour a bug? If not, it does not
seem very obvious to me; I would expect psql to keep track of the
encoding set between the server and the client.

Cheers,

Mathijs