Shane F. Carr
Shane F. Carr
Just like we document structs and enums as either exhaustive or non-exhaustive, we should document traits as either sealed or implementable. For sealed traits, we could add a `__sealed()` function,...
There is a comment in the fallback code: ``` // NOTE: -u-rg is not yet supported; when it is, this test should be updated ``` I think `-u-rg` should be...
See https://github.com/unicode-org/icu4x/pull/4381 We have functions that return both `&ZeroTrie` and `ZeroTrie`. How should we name them to be clear? The current names are not intended to be final.
ICU4X 1.3 started using `~` versions between the metacrate, all component crates, and the data crates, as discussed in https://github.com/unicode-org/icu4x/issues/3537. This solves several problems: 1. Helps prevent CLDR version skew:...
#4204 landed DateTimePatternInterpolator that always loads a FixedDecimalFormatter. Most patterns need one. However, not all patterns do, such as patterns that seek standalone display names. So should we make the...
Currently we have 4 types: - `ZeroTrieSimpleAscii` - Supports ASCII only (so that it can be const constructed) - Does not use the perfect hash function (better for small data)...
The big ZeroTrie PR was merged as #2722 - [ ] Clarify what "bracket a window of strings" means in builder/mod.rs - [ ] Should the `get` function be bounded...
If/when the upstream harfbuzz crate merges https://github.com/servo/rust-harfbuzz/pull/197, we will be able to make icu_harfbuzz be both safe and no_std.
We have "field symbol", which is like "formattable month" (`M`), and "field length", which is like "numeric" or "abbreviated", which together create a "field pattern item" (`MMM`). What do we...
There is `icu::calendar::Date` and `icu::calendar::DateTime` but not `icu::calendar::Time`. We do have a `Time` type but it is in `icu::calendar::types::Time` which is much less discoverable. I think we should re-export `Time`....