I agree with you guys. So, I fixed something in the GlorpNativePostgresDriver so that it gets the encoding info from Glorp's Login and pass it to the PGConnectionArgs :) Cheers, On Wed, Jul 10, 2013 at 12:16 PM, Esteban A. Maringolo <emaringolo@gmail.com
wrote:
2013/7/10 Sven Van Caekenberghe <sven@stfx.eu>:
So...where should I do the conversion? At Glorp level or at database driver level?
And IMHO at the lowest possible level. If you have an SQL VARCHAR column, the driver should convert bytes to a normal Smalltalk String. An OR mapper is a client of the driver, right ?
+1
Encoding should be a parameter of the Connection/Login details. And ORM should deal only with strings in whatever encoding Pharo uses for its strings, ignoring (as much as possible) what happens beneath.
Regards,
-- Mariano http://marianopeck.wordpress.com