Awkward wording of char slicing error message

[steveklabnik]

fn slice_not_on_char_boundaries() {
    let s = "Здравствуйте";
    &s[0..1];
}

gives

thread 'main' panicked at 'index 0 and/or 3 in `Здравствуйте` do not lie on character boundary'

To me, "index do not lie" feels weird, should be "does not lie".

[killercup]

Lol, I initially read this as saying "do not lie [about] character boundary" to the person who wrote s[0..1]!

I'd suggest:

byte index 0 and/or 3 in Здравствуйте is not on a character boundary

Assuming the actual problem is that you can't form a new string slice with invalid/cut off characters:

byte index 0 and/or 3 in Здравствуйте is middle of a character, slice would be invalid UTF-8

[durka]

Well, the problem is "and/or": "0 and 3 are" / "0 or 3 is". Can we just check in the function that prints this message, and make the error message less vague? It's on the panic path anyway.

slice is invalid UTF-8 because byte index 3 is in the middle of char д (bytes 2..4) of Здравствуйте

[bluss]

I made a PR with your strategy @durka. It feels good to be more friendly yet technically specific here (?)