Shane F. Carr

This issue is to track migrating *parts* of until.unicode.org to ICU4X, which has been discussed in various forums, but I couldn't find a canonical issue. Investigation on other server cost...

- @robertbastian Often we don't have the offset/absolute time. You need this for generic zone mapping, too, not just specific zone mapping. - @sffc Maybe it can encode the time...

The 2.0 change should be that VariantOffsetsCalculatorBorrowed should probably operate on a TimeZoneInfo so that it can read the UtcOffset if available.

What we're currently doing is really quite terrible from a datetime hygiene point of view, but the results shouldn't be _unintelligible_ or _outright wrong_, only _misleading_. What we _should_ do...

According to https://tc39.es/proposal-temporal/docs/plaindatetime.html, the default behavior of most datetime libraries is to disambiguate using the offset before the transition. So, if we store local datetimes, we should store them with...

> but if that's what everyone does It's not what "everyone else does" (there isn't really anyone else). These _should_ be absolute times. The only reason we are proposing to...

I want to name it based on what it does versus what it is. `ZoneNameTimestamp`?

ZoneNameTimestamp has a reserved metadata field that could be used to add this. There are still some open questions on improving the data quality here, so I will keep this...

Translators contribute this data to CLDR as a locale-specific setting. Some locales have a space, and others don't. You may wish to search for an issue on [Unicode Jira](https://unicode-org.atlassian.net/) or...

The relevant struct whose construction is being deferred is ```rust /// An adapter between a Unicode back end an the `idna` crate. pub struct Adapter { mapper: Uts46Mapper, canonical_combining_class: CanonicalCombiningClassMap,...