Manish Goregaokar

> Your app needs to be re-written with new assumptions. This is a compelling reason for me to feel more strongly that "allowing a single option to be backed by...

I don't think we need to go as far as to have an icu_util prefix, it seems very inside baseball, but i'm not opposed to it. I like unique names...

My immediate reaction is worry about contravariance and implied bounds sneaking in a `'b` anyway: see https://github.com/rust-lang/rust/issues/106431 for the last time this kind of thing happened. It gets pretty subtle...

`R: 'static` would indeed be a way to make this work without any chance of problems. Would be _nice_ to allow self-borrows, but `R: 'a` is almost certainly going to...

I'm in favor of `R: 'static` + a breaking change

@robofinch does `R: 'static` still work with your use case? We're not opposed to this change, I just don't have the time to do it myself. I recommend making a...

Regarding why you can't get a reference to the cart: you can call `.backing_cart()` on the yoke, no? Ah I understand now. Yeah that's rather tricky to design safe API...

Could you file a separate issue for the "reborrow from cart" use case? I think it's solvable, but I don't think map_project is the way. I'd be more comfortable with...

> I'll make a PR when I have the time, assuming we're still open to the breaking change on Yokeable implementors. @robofinch I'm open to it, but I'd still _prefer_...

> Reading https://github.com/unicode-org/icu4x/issues/2095 earlier today is the first time I heard Pin is currently problematic. I guess I'll need to enable -Zmiri-retag-fields. Feels like a crazy blind spot. Note that...