Le 10 août 2016 20:16, "Steve Dower" <steve.dower@python.org> a écrit :
> So what are your concerns? Suggestions?

Add a new option specific to Windows to switch to UTF-8 everywhere, use BOM, whatever you want, *but* don't change the defaults.

IMO mbcs encoding is the least worst encoding for the default.

I have an idea of a similar option for UNIX: ignore user preference (LC_ALL, LC_CTYPE, LANG environment variables) and force UTF-8. It's a common request on UNIX where UTF-8 is now the encoding of almost all systems, whereas the C library continues to use ASCII when the POSIX locale is used (which occurs in many cases).

Perl already has such utf8 option.