read only one byte in read_single_key unless we find an escape sequence

[rjwalters]

When input is pasted into a unix terminal and collected using read_single_key I noticed that only every 20th character was available. I don't think it should be necessary to read more than one byte at a time. (Hopefully no unicode issues etc.)

[rjwalters]

This causes problems for keystrokes that are longer than one byte... for example in dialoguer's Select functionality. I think we need to look for multibyte keystrokes and read more of the buffer when we find them...

b"\x1b[D" => Key::ArrowLeft,
b"\x1b[C" => Key::ArrowRight,
b"\x1b[A" => Key::ArrowUp,
b"\x1b[B" => Key::ArrowDown,
b"\n" | b"\r" => Key::Enter,
b"\x1b" => Key::Escape,
b"\x7f" => Key::Backspace,
b"\x1b[H" => Key::Home,
b"\x1b[F" => Key::End,
b"\t" => Key::Tab,
b"\x1b[3~" => Key::Del,

[pksunkara]

Sounds good, but how does utf8 work with this?

[rjwalters]

I added special cases to capture UTF-8 based on https://en.wikipedia.org/wiki/UTF-8

[pksunkara]

Thanks. This is great work. Really appreciate it. Feel free to open PRs anytime you want. I will always respond fast. 😄