Rethinking Bundling: Why You Don’t Need Less JavaScript

Tobias Koppers introduces TurboPack and TurboEngine, addressing the limitations of Webpack. He demonstrates live coding to showcase the optimization of cache validation and build efficiency. The talk covers adding logging and memorization, optimizing execution and tracking dependencies, implementing invalidation and watcher, and storing and deleting invalidators. It also discusses incremental compilation, integration with other monorepo tools, error display, and the possibility of a plugin system for Toolpag. Lastly, the comparison with Bunn's Builder is mentioned.

Rome is a toolchain built in Rust that aims to replace multiple tools and provide high-quality diagnostics for code maintenance. It simplifies tool interactions by performing all operations once, generating a shared structure for all tools. Rome offers a customizable format experience with a stable formatter and a linter with over 150 rules. It integrates with VCS and VLSP, supports error-resilient parsing, and has exciting plans for the future, including the ability to create JavaScript plugins. Rome aims to be a top-notch toolchain and welcomes community input to improve its work.

Bun is a modern JavaScript runtime environment that combines a bundler, transpiler, package manager, and runtime. It offers faster installation of NPM packages and execution of package.json scripts. Bun introduces a new JavaScript and TypeScript bundler with built-in support for server components, enabling easy RPC with the client. This allows for code splitting and running code that streamingly renders React or any other library from the server and mixes it with client code, resulting in less JavaScript sent to the client.

TurboPack is a new bundle similar to Webpack, focusing on incremental builds to make them as fast as possible. Challenges in production builds include persistent caching, incremental algorithms, and optimizing export usage. The compilation process can be split into parsing and transforming modules, and chunking the module graph. TurboPack aims to achieve faster production builds through incremental optimization and efficiency. Collaboration and compatibility with other ecosystems are being considered, along with the design of a plugin interface and tree-shaking optimization.

Edoardo, DevRel at Storyblok, explains the importance of JavaScript bundlers and discusses Storyblok's migration to Vite. Challenges with old JavaScript applications are illustrated, emphasizing issues with global variables and dependency control. Optimizing JavaScript module loading through ES modules is discussed, highlighting browser compatibility and performance concerns. The process of creating and structuring JavaScript bundles is detailed, focusing on dependency graphs and module organization. Techniques for managing bundle execution, utilizing abstract syntax trees for code parsing, and implementing optimization strategies are explored, with a specific emphasis on Vite, hot module replacement, and development enhancements.

Tobias Koppers from Vercel TurboPack team discusses implementing persistent caching for long-term web application development, foreseeing significant growth in application sizes over the next decade. AI's role in code writing, TurboPack's vision for instant builds, challenges in build control, and the shift to incremental performance with caching for faster builds. TurboPack emphasizes making incremental builds fast and every build incremental, focusing on trustable and granular incremental builds, efficient granular cache handling in TurboEngine, and automatic granular cache invalidation. The system optimizes performance through efficient persistent cache integration, graph state persistence, database optimization, custom persistent layer creation, and optimizing build time efficiency. TurboPack stands out with its unique bottom-up caching approach, expansion to a general purpose bundler, and plans to enhance framework compatibility. Additionally, the comparison with ESBuild highlights the emphasis on incremental builds and detailed cache granularity with a token-based, almost AI-like approach.