Robert Bastian

I would consider this fixed if we have CI that prevents backsliding.

The `regex` crate apparently used to have a `regex!` macro to compile-time create a regex, but that seems to have disappeared. This can be solved with a custom proc macro...

[`const_regex`](https://docs.rs/const-regex/0.1.0/const_regex/) is prior art, although not directly usable as we need regex-automata 0.2.

`icu_provider_adapters` contains adapters that are providers and wrap providers. This is absolutely not the case in the high-level locale crate, it contains logic that works with locales, so I think...

I'd rather use a shiny new name for the higher-level crate, but won't die on this hill.

* `icu_locid_info` * `icu_locid_meta[data]` A whole new can of worms would be suggesting to rename `icu_locid` to `icu_locale` in 2.0. The crate's main types are `Locale` and `LanguageIdentifier`, there's no...

I don't think ICU4C file names should have any bearing on our naming. Note that the library inside the `icu-locale` crate could still be called `icu_locale`. At some point crates.io...

> and because the name icu-locale was claimed in kebab case and we can't get it back in snake case. I don't think this is a big problem. Most users...

Another suggestion: Most users of a locale type crate will expect there to be canonicalization, minimization, and other locale information. There's a very small use case (basically `icu_provider`) that only...

I guess `icu_core` is fine, although there might be a non-ICU use case for `icu_locale_core` that doesn't require ICU's logging macros. All ICU4X crates (except for locid) already depend on...