Avg Release Cycle
460 days ago
- ⚡️ Wrote rake tasks to update CLDR and ICU resources.
- All resource files now written with symbolized keys so the gem doesn't have to recursively symbolize them on load.
- 🐎 Unicode code points now represented internally with integers instead of strings for better performance.
- ➕ Added telephone code lookup functionality (per country) and postal code validation.
- Collation tries now loaded from marshal dumps, collation running time improved by ~80%.
- ➕ Added case-first collation element tailoring support for languages like Danish.
- 📜 Included a missing development dependency (ruby_parser).
- ➕ Added locale-aware collation via fractional collation element tailoring.
- ➕ Added #sort and #sort! methods to LocalizedArray.
- ➕ Added collation (sorting) support via the Unicode Collation Algorithm.
- ➕ Added Catalan, Basque, Greek, Afrikaans, Ukrainian, and Czech support along with calendar fixes for existing locales.
- 👍 DateTimeTokenizer now falls back on English if the given locale isn't supported.
- ➕ Added ability to use NFC and NFKC in core_ext/string
- ➕ Added NFC and NFKC algorithms.
- 🔨 Refactored Shared::UnicodeData::Attributes into Shared::CodePoint.
- ➕ Added relative time functionality, eg. "2 seconds ago".
- Reorganized locale resources.
- ➕ Added explicit specs for examples in the README.
- ArgumentError now raised if a resource can't be found.
- 🛠 Fixed behavior of the :precision option for number formatting.
- ⚡️ Updated CLDR data to v21 (http://cldr.unicode.org/index/downloads/cldr-21).
- ➕ Added support for localized arrays (i.e. arrays of Unicode code points).
- ➕ Added NFKD normalization algorithm.
- 🐎 Formatter tokens now cached for better performance.
- 👌 Improvements to core extensions (Symbol, Date, etc).
- ➕ Added full normalization test from unicode.org.
- 🐎 Autoload classes to improve performance.