Servo html5ever Parser Runs in Browser via 465KB WASM
Compile Servo's html5ever and markup5ever_rcdom crates to WebAssembly for client-side HTML parsing, handling malformed input like unclosed tags and mis-nesting—full Servo won't compile due to SpiderMonkey, threads, and GL dependencies.
Compile Servo Sub-Crates to WASM for Browser Use
Servo's individual crates like html5ever (HTML5 parser) and markup5ever_rcdom (RC DOM builder) compile cleanly to WebAssembly at ~465 KB, enabling fully client-side execution without server roundtrips. This sidesteps the full servo crate's compilation blockers: SpiderMonkey JS engine, threading, and OpenGL dependencies. Result: a production-grade parser runs in any modern browser tab, producing a parse tree and normalized serialization from raw input.
To integrate, load the WASM module (shows 'loading wasm…' on init), feed it HTML strings, and query the output DOM. Handles real-world mess: input like <p>hello<b>world</p><i>! <div><table><tr><td>implicit tbody <p>unclosed</html><!-- parser adopts the mis-nested tags --> yields Servo's tolerant interpretation—implicit <tbody>, tag adoption despite mis-nesting, and comment preservation.
Interactive Demo Reveals Parser Robustness
Edit HTML live on the left pane; right pane updates with:
- Parse tree: Full DOM structure as Servo sees it, exposing quirks like implicit table elements.
- Normalized HTML: Clean serialization matching Servo's output, stripping errors without losing intent.
This proves Servo components deliver browser fidelity in WASM—ideal for tools needing accurate parsing (linters, preprocessors, editors) without bundling a full engine. Trade-off: sticks to parser/DOM only; no layout or rendering. Start with the demo at https://simonw.github.io/research/servo-crate-exploration/html5ever-wasm-demo/www/ to test edge cases yourself.