Building polyfills in WebAssembly
summary
Claims that ECMA-402 features can ship early by packaging ICU4X as a WebAssembly ponyfill and show concrete results: a grapheme cluster segmenter compiled from Rust with Diplomat and C FFI runs in Node 22 and segments a 100,000‑character string in about 7 ms versus ~35 ms for a pure JS implementation, and swapping ICU for C to ICU4X halved a Skia wasm dependency that once consumed ~20% of the bundle, so the wasm ponyfill is small and fast.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:29Internationalization challenges
- 2:55i18n stack: Unicode, CLDR, ICU, ECMA 402
- 3:59ECMA 402 task group role
- 6:09ICU forx and design goals
- 7:14ICU forx performance benefits
- 8:59ICU forx on servers
- 9:42Polyfills vs ponyfills
- 10:46ICU forx as ponyfill
- 11:30ICU forx with WebAssembly
- 12:35WebAssembly bindings deep dive
- 13:40Live demo: ICU forx
- 16:52Performance vs native ECMA 402
- 17:58Live demo results
- 19:41WebAssembly pain points
- 20:42Conclusion
