Federico Di Gregorio escribió:
Il giorno ven, 10/03/2006 alle 19.35 +0100, Oswaldo Hernández ha
scritto:
[snip]
I forced a conversion of the postgresql error string from latin1 to
unicode and now works fine in the connect moment and after of
client_encoding.
I'm glad you solved it. I'd like to solve this in psycopg too but if
PostgreSQL let you use 2 different encodings (for messages and data) I
don't see any viable solution. :/
The data works fine for me without need of input or output conversion, the problem was only in the
messages.
The last week I spent a lot of time testing encodings between wx, python and
postgresql.
Initially I have python with defalut encoding latin1, postgresql with latin1 and wx unicode version,
there no problem with messages, but I have data problems with some characters ('€' Euro char).
The final composite that apparently works fine whith data was putting UTF8 in
all.
But then ocurr the message problem.
Until now, I ignore the double encoding and i think that it's a big problem.
May be Postgresql should retrieve the messages in the same encoding of the data, converting it if it
be need. Or at least include a new instruction to inform the message encoding, same as 'show
database_encoding' or 'show client encoding'. Without this then programmer never konws how convert
the messages.
Regards
--
*****************************************
Oswaldo Hernández
oswaldo-aXkDUzFGsCcxAGwisGp4zA@xxxxxxxxxxxxxxxx
*****************************************
|
