Unicode input

[khaf]

It seems that the result of get_char does not agree with NCurses::print if they are multi-byte unicode codepoints in NCurses::Key. Key probably needs specific to_char and to_s implementations.

You can simply test this using the examples/input.cr and typing using a non-English input (Easiest would be Russian or Chinese I guess)

I tested on MacOS. Linux is probably the same.

[vllur]

I can confirm that it happens on Linux too. Entering polish diacritics either results in displaying other symbol or nothing.