icu4x
icu4x copied to clipboard
Solving i18n for client-side and resource-constrained environments.
https://github.com/unicode-org/icu4x/pull/2253 adds `offset_fallback`, used when we are unable to display a timezone offset due to it not being specified. Ideal usage of the timezones API will not hit this case...
The ICU4X composing normalizer uses a `UnicodeSet` for a fast-path pass-through check while the ICU4C composing normalizer uses a code point trie lookup. ICU4C ends up being faster ever after...
Currently, `icu_uniset` stores sets as an inversion list and performs binary search on the list. Binary search accesses memory both forward and backward. The [Eytzinger order](https://arxiv.org/abs/1509.05053) orders array elements for...
Related to https://github.com/unicode-org/icu4x/pull/1897 and https://github.com/unicode-org/icu4x/issues/1531. After we move dictionary segmenter to `DataProvider` and testdata, it is better to use GraphmeSegmenter in dictionary segmenter instead.
The tutorials in files such as "intro.md" are woefully out of date. We should make sure they are updated for 1.0, and also automate testing as much as possible, e.g....
Part of https://github.com/unicode-org/icu4x/issues/1082 The zerovec crate has a bunch of operations which may panic, specifically most of the mutation operations as well as VarZeroVec construction operations where the final slice...
We will be able to slightly improve the efficiency of converting ResourceOptions to DataRequest if we only copy over the extensions relevant to a particular data loading operation, which we...
All ICU provider crates should be fully documented. To track this, we should enable `warn(missing_docs)` as suggested in boilerplate.md.
All ICU component crates should be fully documented. To track this, we should enable `warn(missing_docs)` as suggested in boilerplate.md. Pre-existing sub-issues: - https://github.com/unicode-org/icu4x/issues/1025 - https://github.com/unicode-org/icu4x/issues/1028 - https://github.com/unicode-org/icu4x/issues/1029 This issue does...
When we throw away more detailed errors during conversion across FFI, we should first be logging them A tricky thing here is figuring out how the logger gets set up....