Robert Bastian

These are all the remaining undocumented lints that I couldn't fix. Please fix them or tell me how to. * [x] `icu_calendar`, @Manishearth * [x] `Ethiopian::ethiopian_from_fixed` * [ ] `icu_datetime`,...

Actually, `Part`s is also a crate-external link, it's `writeable::Part`. Adding the `dev-dependency` to the meta-crate fixes the `Part` and `Writeable` links, but not the `writeable` link to the _crate_.

I think we should be consistent with ICU4J and ICU4C. Both of them don't seem to have dedicated logos (expect for a terrible logo on the [ICU4J Maven page](https://mvnrepository.com/artifact/com.ibm.icu/icu4j)). I...

What's our policy on icu_datagen supporting old versions of input data? Do we want to support anything after our 1.0 release, forever?

`full-datagen` uses GH Actions infrastructure that we cannot replicate locally. `cargo make full-data` is _almost_ the same, the difference is how the downloading and caching happens (in the binary itself)....

So yeah, just an alias for `full-data`

Roughly: _unstable: This function is unstable in the way that its data provider bounds are not guaranteed to be stable across even minor releases. If you don't need this stability...

Yeah it's not technically dynamic dispatch, but still it runtime-dispatches between many `match` cases, which has a cost, and then has some `Any` overhead as well. > I'd like to...

To clarify I'm advocating against the `icu_provider::impl_auto_deserializing!` macro.

Ah, before my time. I don't think this is consistently used, so let's keep this open to track cleanup?