Simpler transition to PEP 3000 "Unicode only strings"?

["Martin v. Löwis"]

Petr Prikryl wrote:
> Would this break any existing code?

Yes, it would break code which currently contains

# -*- coding: utf-8 -*-

and also contains byte string literals.

Notice that there is an alternative form of the UTF-8
declaration: if the Python file starts with an UTF-8
signature (BOM), then it is automatically considered
as UTF-8, with no explicit conding:-declaration
required. Set IDLE's Options/General/Default Source
Encoding to UTF-8 to have IDLE automatically use the
UTF-8 signature when saving files with non-ASCII
characters.

As for dropping the u prefix on string literals:
Just try the -U option of the interpreter some time,
which makes all string literals Unicode. If you manage
to get the standard library working this way, you
won't need a per-file decision anymore: just start
your program with 'python -U'.

Regards,
Martin