Bundler tree-shaking (and when it lies)
Related service Website & online shop
DEFINITION
A bundler (esbuild, Rollup, Webpack, Turbopack) statically drops unused exports. It works when modules are ESM, exports are unreferenced, and side-effect annotations are correct (`sideEffects: false` in package.json). It lies when you consume CJS, use dynamic `require` / `import`, or the library leans on `import './polyfill'` patterns. Most 'I expected this bundle to be smaller' moments are this · inspect with `webpack-bundle-analyzer` or `source-map-explorer`.
- SSR (Server-Side Rendering)→
HTML rendered by the server per-request, fresh for every user. Ideal for dynamic content (dashboards), but slower than SSG.
- SSG (Static Site Generation)→
Pages are produced at build time as HTML and served from a CDN. Near-zero TTFB. DField's own site runs this way across 111+ pages.
- ISR (Incremental Static Regeneration)→
SSG + timed regeneration: the HTML is static but regenerates on a schedule. Ideal for blog content · freshness with CDN speed.
- Edge rendering→
Code runs at the CDN edge closest to the user (Cloudflare Workers, Vercel Edge). Dynamic responses with ~10–50 ms TTFB.
- RSC (React Server Components)→
React components that run exclusively on the server and never ship to the browser. Result: less client-side JS and faster hydration.
- LCP (Largest Contentful Paint)→
Time until the largest visible element paints. Google Core Web Vitals passes under 2.5s · we usually land marketing pages under 1s.
- 0126 Apr 2026Postgres BRIN vs. B-tree · when each wins→
- 0226 Apr 2026Server vs. Client Components in 2026 · the rule we apply→
- 0323 Apr 2026Bun 1.2 in production · what we keep, what we skip→
- 0422 Apr 2026Speculation Rules API in 2026 · near-instant nav, zero JS cost→
- 0522 Apr 2026pgvector at 10M+ rows · index choice, query patterns, real performance numbers→
- 0622 Apr 2026LLM prompt caching in production · a 60-80% cost cut→
- 0714 Feb 2026Core Web Vitals on Next.js: from 4.5s LCP to 0.9s in 3 weeks→