Hm, interesting, I would have thought it’d be faster to download the cache than pulling from npm. Our node_module’s cache step is 15s and pnpm install step is 10s in CI atm, also a monorepo.

Is that with node module caching?

😨

This fix for this is so slick: instead of reverting back to the old behavior where Suspense fallbacks have to wait for suspended siblings to render, the new feature commits fallbacks immediately, then schedules a render for the siblings after to "pre-warm" the lazy requests in the tree.

swift = hipster ocaml
rust = systems ocaml
elm = baby ocaml
sml = boomer ocaml
f# = enterprise ocaml
rescript = web ocaml
kotlin = jvm ocaml
coq = nerd ocaml
gleam = shiny ocaml
scala = drama ocaml
haskell = lazy ocaml

just learn ocaml already

Want: opentelemetry instrumentation for ava tests

Is it possible to integrate it into an existing backend, a bit like a backend for frontend, like you can do with graphql via resolvers? We already have realtime events going and everything! Or is it straight to Postgres? Which would bypass a lot of the business logic, authz, etc.

Zero_-_20241001 This is "Zero_-_20241001" by Aaron Boodman on Vimeo, the home for high quality videos and the people who love them.

Sneak peek of what we've been up to on Zero:

vimeo.com/1022664777?s...

Zero is the next iteration of Replicache. It's how we're bringing local-first apps to the mainstream.

If you're building a local-first app this year reach out. We're onboarding people in private beta now.

Want a smaller font on the @bsky.app! It’s open source, isn’t it 🤔