Next.js 15.3
• Turbopack for builds (alpha)
• Community support for Rspack (experimental)
• Client Instrumentation hook
• Navigation hooks
• TypeScript plugin improvements
nextjs.org/15-3
Next.js 15.3
• Turbopack for builds (alpha)
• Community support for Rspack (experimental)
• Client Instrumentation hook
• Navigation hooks
• TypeScript plugin improvements
nextjs.org/15-3
You can pass sub-classed Promises to React such as in use() with the fields status and value or reason.
This allows React synchronously read the value without waiting on a microtask. This is much faster but it also ensures compat when someone needs flushSync().
Microtasks are bad, mkay.
don’t useEffect, but if you must, don’t call setState in it
Screenshot of the Vercel dashboard showing graphs for GB-hrs saved, cold and hot starts for functions, and the time to first byte.
Two things I'm excited about at Vercel right now:
1. Saving developers $$$
2. Improved observability
Turn on in-function concurrency, redeploy, and save ~40-70% on usage.
Plus, more info on cold/hot starts, initial load performance, and more coming soon.
vercel.com/changelog/in...
My talk from #nextjsconf is now available on YouTube!
Having poor performance with Partial Prerendering with @nextjs.org? Check your code if you're using revalidatePath. This is actually revalidating the entire static shell each time, making things really slow! Instead rely on router.refresh() to update your server component data.
we’re also gonna introduce a new refresh() api to do this directly from a server action in 15.1 instead of doing this at the client level
Random insight: The website was slower because of the theme picker. Why?
The site had a color and background transition (not GPU accelerated) on the theme picker for when a selection is made.
🧵
you can just ship things. so i did.
if you have a domain on vercel, it's now even easier to set up your 🦋 handle.
Excited to support Bluesky’s growth. Domains ftw!
What are all the reasons "real user metrics" for Core Web Vitals would differ from Google's PageSpeed Insights? I'll start:
- RUM data may include non-Chrome browsers (For LCP and INP they would all be Chromiums)
- Google only counts users who have Chrome sync on and opted into data sharing
OSS