The dependency unic-ucd-ident has not been updated in 5 years upstream. Since then there have been at least some small changes to ID_Start/ID_Continue. For example U+30FB KATAKANA MIDDLE is not handled correctly. I found https://github.com/oxc-project/unicode-id-start, which is a fork of https://github.com/dtolnay/unicode-ident (discussed in dtolnay/unicode-ident#1). I had hoped that unicode-ident would just support this.