diff chunks for certain unicode edits incorrect

[kitsonk]

I am getting unexpected result with diff() when using some unicode chars. Specifically I think the problem is with the 🇺🇸 which is a combinatory sequence of combining 🇺 Regional Indicator Symbol Letter U and 🇸 Regional Indicator Symbol Letter S.

let a = " 🦕🇺🇸👍 ";
let b = " 🇺🇸👍 ";
let actual = diff(a, b);
assert_eq!(actual, vec![
  Chunk::Equal(" "),
  Chunk::Delete("🦕"),
  Chunk::Equal("🇺🇸👍 "),
]);

Which fails with:

thread panicked at 'assertion failed: `(left == right)`
  left: `[Equal(" "), Delete("🦕🇺"), Equal("🇸👍 ")]`,
 right: `[Equal(" "), Delete("🦕"), Equal("🇺🇸👍 ")]`'

I haven't dug into the code to see why it is breaking "incorrectly" across the unicode sequence, but wanted to raise the issue first.

[derekperkins]

Looks like this is also an issue with the original lib? google/diff-match-patch#59

[dtolnay]

Fixed in 1.0.6.