It's better but still not good

My table collocation is latin1_general_ci (all _dbtech_* tables).
After update I don't have special characters like "?" but I see "raw" characters from database.
Now looks like that
http://imageshack.us/photo/my-images...bshoutbox.png/
Some characters: n = ? (it should look like this n = ń) ; l = ? (l = ł) ; c = ? (c = ć) ; a = ? (a = ą) etc.
This the way how polish characters are stored in the database with latin1_general_ci encoding . In older version all works fine (under 6.x).