Generally speaking browser submisisons were/are supposed to be sent using the same encoding as the page, so if you're sending the page as "latin1" then you'll see that a fair amount I should think. If you send it as "utf-8" then you'll get 100% utf-8 back.
The only trick I know is to use <meta charset="utf-8">. Would that 'send' the post as utf-8? I always expected it had more to do with the way the user entered the characters. How do they by the way, enter things like Montréal, Quebéc. When they enter that into a text box on a web page can we say it's in a particular encoding at that time? At submit time? -- https://mail.python.org/mailman/listinfo/python-list
