David Strawn

@zarthross @zmccoy The pointer tests failed. I have a vague memory of them just being generally flaky. Is that right?

Tests are expected to fail right now. This branch currently still has the tests I added for CIString to debug/validate the full case folding behavior on CIString. Since CIString is...

> If CIString is deprecated anyway, why not just let it retain its old implementation, for sake of compatibility? Unless I misunderstood and that is what it was doing. My...

@rossabaker > This seems like mostly a superset of the other one, so it's probably better to discuss the design there. This one drops the generic `CaseFoldedString` for separate types....

@rossabaker @armanbilge I've pushed a new commit which undeprecates and changes `CIString` to be Unicode canonical caseless matching with full case folding as discussed [here](https://github.com/typelevel/case-insensitive/pull/232#discussion_r800251110). ~See: https://github.com/typelevel/case-insensitive/pull/232/commits/b376e37a65ebd43fac9b8007c9f30dbd6c790fa3~ I ran into...

Yeah, it's not great. I'm still open to suggestions on what to do here. We could provide an alternative to to the `ci` syntax, but still not great.

My concern with not changing this until http4s 1.x.x is that what CIString does right now is (probably) equivalent to simple case folding with Turkic rules and default caseless matching....

@rossabaker I'm (finally) circling back here. Do you have any thoughts? I suspect there are some very deep serious side effects of the incorrect case-folding, but that they are quite...

> Eyeballing this after several months away, it seems like it's binary compatible, and I didn't set off any MiMa alarm bells. Am I correct that this a purely semantic...

@armanbilge Do you just mean `CIString` encoded as an opaque type or are you talking about wrapping `CIString` again? And are you foreseeing us having `hashCode` differ from `Hash`?