The Weekly Dev's Brew #26 ☕

Oxlint Finally Brews Type-Aware Linting

The folks at Oxc just dropped the type-aware preview for oxlint, and it's bringing some serious performance improvements to the TypeScript linting game. The headline numbers are impressive: repositories that previously took a minute now complete in under 10 seconds with their new setup.

The secret sauce? They're leveraging typescript-go – TypeScript written in Go that's 10x faster than the original. The architecture is clever: oxlint handles the frontend (CLI, file traversal, diagnostics) while tsgolint acts as the backend, creating a simple but effective pipeline.

What makes this particularly interesting is their approach to the type-checking challenge. Instead of writing their own type checker (which has been attempted many times with mixed results), they're accessing TypeScript's internal functionality that wasn't originally meant to be used by external tools. It's a bit like using a library's private methods – not officially supported, but it gets the job done.

Now, before you get too excited, this is still a preview release. The team is upfront about current limitations: large monorepos might hang, there are some stability issues, and you can't configure individual rules yet. But the foundation is solid, and they're actively addressing these issues.

What I appreciate most is their transparency about the technical decisions and roadmap. The v1.0 plan includes performance improvements for large monorepos, individual rule configuration, IDE support, and overall stability. If they deliver on these promises, this could be a compelling option for teams looking to speed up their TypeScript workflows.

Podcast Schedule Update 🎙️

We're moving the podcast from Mondays to Fridays! This gives me a bit more breathing room over the weekends (because who doesn't need that extra cup of coffee on Sunday morning?). This Friday, I'm sitting down with Dominik Dorfmeister, TanStack Query maintainer and Sentry's newest team member (more or less). We'll be diving into his journey with TanStack Query and what his new role at Sentry looks like. Don’t forget to subscribe, you don’t want to miss this!

Quick Sips

• CSS-Only Scrollspy Magic: Sara Soueidan dropped an excellent deep-dive into CSS-only scrollspy using the new scroll-target-group property and :target-current pseudo-selector. You can now highlight navigation links when their target sections are in view without JavaScript. Chrome 140+ only, but it's a glimpse into more declarative UI patterns.

• Biome 2.2 Upgrades Import Management: The noRestrictedImports rule now supports gitignore-style patterns for restricting module groups with exceptions. Plus, new sorting customization for imports, keys, and attributes with natural vs lexicographic ordering. Several rules graduated from nursery to stable, including noAwaitInLoops and useConsistentObjectDefinitions.

• Learn Signals the Right Way: Jon Kuperman created a minimal, educational implementation of JavaScript Signals based on the TC39 proposal. It prioritizes clarity over performance – perfect for understanding the concepts without getting lost in optimization details.

• Complete Svelte 5 Reference Guide: Need a comprehensive resource for Svelte 5? Matia wrote the complete reference guide, that that covers everything from runes to the new reactivity model. Perfect for developers making the transition or just getting started with Svelte's latest iteration.

Coffee Fact of the Week ☕

The "espresso" brewing method was invented in 1884 by Angelo Moriondo in Turin, Italy. The word comes from the Italian "esprimere," meaning "to express" or "to press out" – which is ironically what we're all trying to do with our code when deadlines approach. The pressure-based extraction method produces coffee in about 25-30 seconds, making it the original "fast build" of the coffee world.

See you next week for more caffeinated web dev news!

The Weekly Dev's Brew - Your Morning Companion for Web Dev Insights