Re: Charset encoding patch to JDBC driver

[Oliver Jowett <oliver(at)opencloud(dot)com>]

Javier Yáñez wrote:
> I have the necesity of to keep a PostgreSQL database with SQL-ASCII. As 
> the actual version of pgjdbc only have support for Unicode (at least for 
> jdbc3), I have make a patch that allow to configure the desired charset.

I'm uncomfortable with applying this sort of patch to the official 
driver, since it makes the driver more complex just to handle what is 
arguably a database misconfiguration. It also introduces a new class of 
error: a mismatch between the driver's configured charSet and the actual 
database.

Comments on the patch itself:

- it is missing changes to the v2 protocol path
- why does it remove the client_encoding sanity check on connect?
- since encoding does not change for the lifetime of the connection, 
can't you make the encoding a field of QueryExecutoryImpl rather than 
passing it around everywhere?
- it may be better to pass encoding as a parameter to 
SimpleParameterList methods that need it, rather than storing the (same) 
value on every list instance.

-O