Should U+22A5 and U+27C2 be equivalent?

[jonas-schulze]

This is a follow up to JuliaLang/julia#39350 (comment).

In Julia U+22A5 (⊥ or \bot) and U+27C2 (⟂ or \perp) are different, even under normalization. Is this intended behavior or should they be equivalent?

cc: @mcabbott, @StefanKarpinski

[StefanKarpinski]

The Unicode Consortium decides on normalization rules, this repo just implements them. That said, Julia does some additional normalization, IIRC.

[stevengj]

Yes, they aren't equivalent according to Unicode, so this is not a utf8proc issue.

See also the discussion in #11: we do provide the hooks for custom normalizations, and Julia takes advantage of this (JuliaLang/julia#19464). So, we could add it as a new Julia-specific normalization, but we've been conservative about this — basically, we've used it for confusable characters which show up frequently because they have popular input methods.