TL;DR
To implement confusable_idents lint in RFC 2457, we need a mechanism to collect all unique identifiers occurred during compilation.
Links and Details
The requirement raised up from RFC 2457. Fundamentally, it means that each occurred identifier should be hashed with a very specific hash function called skeleton and storing the results. The lint is about checking whether hash collisions occurred during compilation. Whenever that happens, it raises warnings or errors.
Mentors or Reviewers
(TBD)