util::split_at panics with multibyte characters

[kyoheiu]

Sorry if duplicated.

To reproduce panic, run the following code:

use syntect::easy::HighlightLines;
use syntect::highlighting::{Style, ThemeSet};
use syntect::parsing::SyntaxSet;
use syntect::util::{as_24_bit_terminal_escaped, split_at, LinesWithEndings};

fn main() {
    let ps = SyntaxSet::load_defaults_newlines();
    let ts = ThemeSet::load_defaults();

    let syntax = ps.find_syntax_by_extension("rs").unwrap();
    let mut h = HighlightLines::new(syntax, &ts.themes["base16-ocean.dark"]);

    let s = "日本の首都は東京です";
    for line in LinesWithEndings::from(s) {
        let ranges: Vec<(Style, &str)> = h.highlight_line(line, &ps).unwrap();
        //thread 'main' panicked at 'byte index 4 is not a char boundary; it is inside '本' (bytes 3..6) of `日本の首都は東京です`', library/core/src/str/mod.rs:127:5
        let ranges = split_at(&ranges, 4);
        let escaped = as_24_bit_terminal_escaped(&ranges.0, false);
        print!("{}", escaped);
        println!();
    }
}

I think it would be better to return Result<(Vec..., Vec...), Error> instead of (Vec... , Vec...) to avoid this type of panic.

[kyoheiu]

Seems this panic comes from this line:

        let (sa, sb) = rest[0].1.split_at(rest_split_i);

[str - Rust](https://doc.rust-lang.org/std/primitive.str.html#method.split_at)

Panics if mid is not on a UTF-8 code point boundary, or if it is past the end of the last code point of the string slice.

So it may be resolved by using is_char_boundary() ([str - Rust](https://doc.rust-lang.org/std/primitive.str.html#method.is_char_boundary)), but depends on how you think about this.

[Enselic]

Thanks for reporting. Would be great if you could create a PR with a proposed fix together with a regression test.

[kyoheiu]

Created 2 PR to fix this issue: One is to fix this function to return Result, another to return the same type as before, just decreases the index to avoid panic. Which is preferred I cannot say, so please check.