How Small Can You Go? Customizing Interpreted Languages for Wasm Environments
Talkat Cloud Native Wasm Day Europe 2023 · on Tuesday, 2023-04-18· at 16:20· Hall 7, Room C | Ground Floor | Europe Complex·
summary
Explains how interpreters for Python, Ruby and PHP shrink for edge Wasm by trimming dependencies, slicing the standard library and offering configurable builds; demonstrates slim runtimes ~5–13MB and a mid‑build with XML+SQLite at ~10MB. Shows a prototype ycvfs compression that cuts an embedded WordPress Wasm from 71MB to 29MB and a workflow to link tiny user modules (even 300 bytes) into a cached interpreter via pre-open, stdin or module linking. Notes current decompression is slow but functional and advocates WASI/component APIs to replace bulky stdlib features.
AI-written summary · about
chapters (7)
- 0:00Start
- 0:15Why: Interpreted languages
- 1:58Interpreter based optimizations: size
- 4:05Standard library based optimizations
- 5:50Interpreter + Application
- 13:40Providing Source via Module Linking
- 20:51Building bridges
