Good point.

So to avoid encoding bugs and warnings at output about "Wide character in print" one can write:

use utf8;
$ENV{LANG} = 'ru_RU.UTF-8';
use open IO => ':locale';   # I tried ':utf8' instead but got warnings
[download]

The bugs to which you linked are Encode's, not encoding. While they may effect use encoding, I was referring to use encoding's proper flaws. Now, I can't remember exactly what they were, but you can search the P5P mailing list for the conversation.

Depending on what you want to do,

# Use expected encoding for STDIN/OUT/ERR
use open ':std', ':locale';

# Use UTF-8 for STDIN/OUT/ERR
use open ':std', ':encoding(UTF-8)';

# Use UTF-8 for file IO
use open IO => ':encoding(UTF-8)';
[download]

:utf8 has a security vulnerability for input.