Manish Goregaokar

> I wanted to discuss the ZeroVec iterators because they are currently written as Ah, that's a different problem than `impl Trait`. I would measure this. 99% of the time...

0. Live with it (`icu_timezone`, `icu::timezone`, `TimeZone`) 1. Change `icu_timezone` to `icu_time_zone` and deprecate the crate `icu_timezone` 2. Change `icu::datetime::time_zone` to `icu::datetime::timezone` in module names 3. 2 + change all...

I also vote 2 2 > 3 > 0 >> 4 > 1

The repr doesn't matter since C++ will be converting after C anyway.

This should be filed on Diplomat as a special case for C++. I don't know the best way to do this, I'm reluctant to add a Diplomat type system variant...

I'm inclined to say that we are not doing this since this is a bunch of Diplomat work for a hyperspecific, C++-20-only type.\ If someone wants to do the work...

I'm fine with this policy with the caveat that updating patch releases is fine. Basically, if 1.1 doesn't build on 1.69 but if performing a _patch-only_ lockfile update (1.1 ->...

> allowing patch updates does not enable us to test the world at arbitrary points in time I think to some extent "the world" is not static, anyone using software...

Yeah, we added `ordinal` because it might be useful for users, but it's not useful for formatting. We need to disunify formattable month and "regular month info a user might...

We should redesign these methods to be Month/Year _by default_ with the formattable ones being separate methods