micromark@4.0.2 · LilScript port · @itslil/micromark
Markdown,
compiled
This is not micromark. It is a LilScript markdown→HTML compiler whose 660 CommonMark+GFM cases match marked@18.0.10 HTML. Official rows are the published micromark@4 runtime graph, then Terser and esbuild.
npm install @itslil/micromark
Official rows are one esbuild bundle of micromark@4.0.2 plus the packages it needs at runtime, then Terser or esbuild.
LilScript is not post-minified. The npm file is the Brotli-scored compile. Closed LilScript is the same source with extern_fields = false.
Measured in Node
Same work,
two runtimes.
Not Playwright — Node v22.21.1 on this machine, quiet median after warmup. Race 10× in the playground for a browser number. The 660-case corpus is checksummed before any size row is a claim.
| Lane | median | vs official |
|---|
The compiled runtime, in the browser
Live
lab.
Left is the editor. Right is @itslil/micromark HTML, or the official micromark@4.0.2 graph when that lane is on.
The receipts
Raw, gzip,
and Brotli.
Official micromark@4.0.2 bundled with its runtime graph, then Terser (mangle on/off) and esbuild minify.
LilScript is not post-minified. Closed LilScript is Brotli with extern_fields = false.
Brotli-11
The wire number. LilScript row is the npm file.
| Lane | Brotli-11 | vs Terser |
|---|
gzip-9
Same files, gzip-9.
| Lane | gzip-9 | vs Terser |
|---|
Raw
Uncompressed bytes.
| Lane | Raw | vs Terser |
|---|
All three
Every lane, measured three ways.
| Lane | Raw | gzip-9 | Brotli-11 | Brotli vs Terser |
|---|
Why this surface. Comparing against a single published file would hide official dependencies. The fair fight is the runtime graph you actually import, then the same minifiers the JS ecosystem ships. Closed LilScript is the other honest comparison: no JavaScript options object, so those property names mangle.
README ↗Zero runtime dependencies
Install the
current port.
npm install @itslil/micromark