Shane F. Carr
Shane F. Carr
Many, or most, of our functions that are named `from_bytes` are fallible. @markusicu suggested that we follow the Rust convention of `try_from_...` to indicate a fallible function. Thoughts? - [...
A general DataBake feature that would solve several problems, including #2417, would be a way to specify a different data struct as the serialized form of the main data struct,...
Feedback from @markusicu, which I very much agree with, is to rename `icu::locid::extensions::private::Key` and `::other::Key` to `Subtag` instead of `Key`. Markus says that the spec does not make any statements...
This PR is based on the excellent feedback on https://github.com/unicode-org/icu4x/issues/2348#issuecomment-1249563656. I made the following improvements: 1. A brand new documentation page about constructors, which includes a decision chart similar to...
@markusicu had the following suggestions for icu_locid docs that don't affect the API surface: - [ ] In various functions that take UTF-8 bytes, change it to say "ASCII bytes"...
They are internal things, except for potentially removing the `DataKeyHash::to_bytes` API (which I prefer to keep). See #2565
Based on extensive discussion surrounding #2506, we decided to move forward with a design that adds `DataProvider` in `LocaleFallbacker::try_new_unstable`, meaning that this key is unconditionally matched when fallback is used,...
Feedback from @markusicu: > Why is bidi its own module? It's odd that it's inside of properties. It seems like it is higher-level than properties. > > I see, it...
We need bytes-based parsing APIs for FFI compatibility. We should add one for CustomTimeZone. See #2320
Previous issue: https://github.com/unicode-org/icu4x/issues/1226 We need to ensure that the case mapping crate conforms to the ICU4X style guide, including documentation and FFI. CC @iainireland to add anything else that we...