Manish Goregaokar

I disagree with the commit changing the bounds of the Chinese calendar. Firstly, the calendar is not that old (even the thing about January 20 in 30 AD is dubious,...

It's no longer an optimization since we have unused space there but I'd still prefer to keep it around, if the computer is producing dates outside of that range the...

Ah, I was reviewing this on a phone and couldn't see all the comments. I am still leaning towards considering those as cases we shouldn't try to fix. I'm somewhat...

> According to Wolfram Alpha ([query](https://www.wolframalpha.com/input?i=winter+solstice%2820000+BC%29+in+UTC)): I understand this. I don't think that affects what I said: These dates are beyond precise predictability and it is not meaningful to talk...

Right, I understand your proposal, I don't want to increase the stack size of the Date type here. I think the status quo with an explicit clamp for the solstice...

Two other options that came up during discussion: 4. `ZeroBox`: We introduce a type `ZeroBox`/`VarZeroBox` which has the semantics of a boxing `Cow` (by default Cow goes to the `ToOwned`...

Related: https://github.com/unicode-org/icu4x/issues/3020

- @sffc I think it's a no brainer to try specifically fixing DateTimeFormat on the pre-2.0 timeline. Fancy cross-cutting things like ZeroBox/etc are maybe not worth it - @Manishearth Do...

Issues filed: 1. https://github.com/unicode-org/icu4x/issues/3599 2. https://github.com/unicode-org/icu4x/issues/3598 3. https://github.com/unicode-org/icu4x/issues/3347 (existing issue) 4. https://github.com/unicode-org/icu4x/issues/3600

I don't think so.