A bug with a cat and a bat and javascript and fe0f (unicode variation selector)

[th1000s]

Highlighting the string "`😸️" (which is an apostrophe followed by a cat emoji with a variant selector to indicate color) using javascript syntax works fine when using default syntect settings.

However when using bat assets (bat::assets::HighlightingAssets::from_binary()) the variant selector is split off: "😸\u{fe0f}" vs "😸", "\u{fe0f}". Further downstream this confuses delta, see dandavison/delta#1172. Building or clearing the bat cache has no influence on that.

Since my understanding is that no syntax definition should be able change these low level unicode bits I am opening an issue here and not with bat.

See the example HighlightLines::highlight_line() call in batcatbug.rs or to reproduce clone https://github.com/th1000s/bat and run cargo run --example batcatbug with branch highlighter_bug.