Rspack 2.2 Lands With 30+ Performance Patches and Solid 2.0 Support
The HMR fix alone removes a scaling penalty that made hot reloads slower as CSS grew, a pain point for any large app. Combined with the Node.js chunk-splitting defaults in Rsbuild, teams get smaller SSR deployments and lower memory pressure without extra configuration.
Rspack 2.2 delivers more than 30 targeted performance improvements, cutting production build times by roughly 5% and speeding up CSS parsing by up to 3.7x. A new `compat-hashed` module ID strategy trims bundle sizes by picking the shortest stable hash prefix, while HMR no longer fetches unchanged CSS, making JavaScript hot updates consistently fast regardless of stylesheet weight.
Rsbuild 2.2 arrives alongside it with default chunk splitting for Node.js builds, which one TanStack Start user measured as a 98% reduction in server artifact size and a 73% drop in memory usage. Solid 2.0 RC support is also in, switching to Solid's Rust compiler for a claimed 20x compile-speed gain over Babel.
The broader Rstack ecosystem sees Rslint pass 500 built-in rules with full `@typescript-eslint` parity, Rstest add Module Federation and Playwright E2E testing, and Rslib adopt TypeScript 7 for 5–10x faster type generation. An Agent Plugin based on Agent Plugins 1.0 lets Copilot and Cursor pull in Rstack-specific skills directly.
Decoupling HMR latency from CSS size is a structural fix, not a micro-optimization; it eliminates a regression that grew silently with every new stylesheet.
The `compat-hashed` ID strategy is a low-risk, zero-config way to reclaim kilobytes from production bundles without changing caching semantics.
Rsbuild's Node.js chunk-splitting default turns on a capability that many SSR setups leave off, and the 98% artifact-size reduction suggests widespread under-configuration in the wild.
Solid 2.0's Rust compiler landing in Rsbuild signals that framework-level Rust tooling is moving from experimental to default, which raises the bar for Babel-based alternatives.
Rslint reaching full `@typescript-eslint` parity removes the last major reason to keep ESLint in a Rstack pipeline, consolidating the toolchain further.
The Agent Plugin play is a distribution move: embedding Rstack-specific skills into AI coding agents lowers the friction for new users who encounter the toolchain through Copilot or Cursor.