Devon Govett

Introducing the new React Aria docs! 🎉 All-new content and search experience. Interactive CSS and Tailwind examples to get you started quickly – just install with shadcn. New guides and full example apps. MCP server and AI integrations. Check it out! react-aria.adobe.com

Small update to the Skeleton component. You can now wrap your actual UI with Skeleton, set isLoading, and everything inside will automatically animate as a loading state. Cleaner API, less duplicated placeholder markup.

@ForceTheIssue GitHub issues

@devongovett is there a preferred place/method for RAC feature requests? I know you’ve asked on X in the past and there are some issues/discussions on GitHub. We’ve got a few to propose and could even possibly contribute.

@boshen_c There was a recent update about it #issuecomment-4328422631" target="_blank" rel="nofollow noopener">github.com/facebook/react…

The progress of the official React Compiler Rust port seems stale. Should we attempt our own Oxc port and take ownership of its maintenance? We will need a few more external core contributors to keep this thing sustainable.

Finally had time to write up some thoughts on what I need help with and the project philosophy: github.com/parcel-bundler…

This is incredible, but to be honest, I could use some help with maintenance. I appreciate sponsorships, but unfortunately this does not magically create more time for me to work on it. CSS is evolving very quickly and it's difficult to keep up with all the new features as a solo dev. I'm sorry if my slowness has affected you. Now that Lightning CSS is a core ecosystem dependency, I need to find a way to improve this. I'll be reaching out to each of these tools soon to try to bring on representatives as maintainers. And if you're interested in helping, please reach out as well. Hopefully we can get a good core team established and speed up the process a bit.

Looks like Vite 8 is using Lightning CSS as the default CSS minifier! Most next-generation build tools are now using it, including Turbopack, Rspack, Tailwind, Parcel, Bun (a port), and now Vite. ⚡️

@threepointone Still is Sunil! 😅

remember when programming was free

@vpetryniak @ericclemmons Yeah we made a separate submitAction prop for our Form component that doesn’t reset it.

@devongovett @ericclemmons Of course it hasn’t, why anyone would adopt such buggy API ? But that’s just my point of view and maybe there are other reasons. One of the things: github.com/facebook/react…

<LoadingButton /> No `isLoading`. No `useState` in parents. yield your `onClick` handler. stackblitz.com/edit/ericclemm…

@vpetryniak @ericclemmons It hasn’t been well adopted by component libraries yet. I’m working on it for React Aria. Will handle errors too.

@devongovett @ericclemmons Honestly action props is the worst and most buggy api in react. Bugs stay open for years, and UX is atrocious… Errors in action go to error boundary which will discard all user input UI… The tanstack-query useMutation is much better.

@ericclemmons Well maybe not the “taking longer than usual” part, but it’ll handle the loading states for sure

@devongovett All of this? 🤨 Since this did some numbers, there’s a less contrived version where this shines. But it means I need to use effect ;)

Omg it actually happened! 🤣🤣🤣 github.com/oven-sh/bun/tr…

lol will be hilarious if this ends up porting lightningcss -> zig -> rust (but slop) 😂

@colmtuite @diegohaz We build it in so that developers don’t have to think about it. They implement drag and drop and it just works for keyboard users too. They don’t even need to change their UI to support it (eg adding a menu). The menu approach is quite limited and slow to use anyway.

@diegohaz Yeah RA does something like that. I too have mixed feelings about it. So far I think it's a mistake. I'm not convinced they're connected. Keyboard support is just mandatory, and is the responsibility of the higher-level component. I think perhaps DnD can be entirely optional.

i can't stop

React Aria has built-in support for drag and drop. This kanban board is only ~100 lines, and even supports keyboard accessibility with no additional code needed! 🙂

@boshen_c We're also using oxfmt. Nice work!

Oxlint -> React Aria

@thiago_peres I used their eslint migration skill. Only took a few minutes. #migrate-with-skills" target="_blank" rel="nofollow noopener">oxc.rs/docs/guide/usa…

@devongovett nice! i use it in a small project, time to try /goal overnight hehe

Tried out oxlint on the React Aria repo: eslint: 78.13s oxlint: 35.27s oxlint without react-compiler rules: 2.9s Can't wait for the React Compiler Rust port. 😅

@thiago_peres yup, and including our custom eslint plugins unmodified.

@devongovett oxlint with exact same rules?

@theorangenotice does it support running eslint plugins? we have a lot of custom rules that we need to enforce. oxlint worked with them without any changes which was nice.

@devongovett Have you tested biome? It's fast

@royalicing Mmm, no I don’t agree with that. Concurrent rendering is not about performance.

Yes, I get the sense that of the recently added hooks like useTransition that even the authors haven’t fully understood how they should compose with the rest of React. All to support concurrent rendering that most React apps don’t need, and which should be solved by different paradigms or optimization.

Creating your own slop forks of battle tested libraries to make them smaller is a great way to introduce subtle bugs for edge cases you didn’t know you had, create accessibility issues you didn’t know to look for, security problems, etc. The reason we use libraries has never been just to avoid the cost of writing the code ourselves. It’s to start from a baseline level of quality instead of discovering all those edge cases in production.

@royalicing Definitely agree that react should be made more modular.

@devongovett I agree, but I think specifically today’s React can be simplified with a reduced featureset for what many apps actually need. These slop forks can help highlight what particular features that original library brings that are actually necessary vs not.

@gill_kyle Way better to do it in those libraries so everyone can benefit than creating separate security issues in every product that are much less likely to be found or fixed.

@devongovett while I think this is true a lot of the time, I'm not sure it's true 100% of the time not every library and dependency is in its perfect final form or we wouldn't be patching CVEs in the most used libraries on the web every week