**Introducing**.The easiest way to use Paraglide in NextJS to build an internationalized App. 💨 A breeze to set up - no need to change your `app/` folder.📦Built for RSC .🪄 Supports both App & Pages router.🤖 SEO friendly out of the box. Info below 👇

This will be the year of the linux desktop.

Generate videos in just a few seconds. Try Grok Imagine, free for a limited time.

RT @ReactSummit: Great insights from @loris_sigrist on #React Server Components for improved internationalization solutions. #TechConferenc….

This is useful if you.- want to rework a package, but don't want to update every consumer upfront. You can temporarily have both versions of the package to ease migration.- want to be able to publish (security)-patches to legacy versions of your package.

Packages in your workspace can import any version of the package using `workspace:<semver-range>`

pnpm allows you to have multiple packages with the same name in a workspace. Use this to keep multiple versions of your package "alive". When you start work on a new major version, copy the package folder & bump the version number. Then work in the new folder.

Have you ever wanted to rework a package in your monorepo but still want to keep the old version around for patches while you migrate?. If you're using pnpm workspaces, do it like this 🧵.

RT @r0derik: @pontusab @nextjs recently use the toolkits from @inlangHQ , was a really pleasurable experience.

RT @paraglidejs: The upcoming version of Paraglide-Next adds support for typedRoutes. Get autocompletion on Links and useRouter https://t.….

Anyways, JS can be goofy, I just have to deal with it.

You could argue this makes sense since a raw buffer doesn't have an "encoding". However, in most people's mental model "encoding" just means "interpret the data like this". In that sense, "no encoding" still constitutes as an encoding; the identity encoding.

If you want `readFile` to return a Buffer with the raw binary data you need to omit the second argument entirely

Why? . `encoding: "binary"` is an alias for `encoding: "latin1"`. "latin1" is "binary" because there are 256 characters in so it uses 1 byter per char. This makes some sense from a historical standpoint, but today it just seems ridiculous since everyone defaults to utf-8.

I don't really care for most weird JS type behavior but this one is just vile

I don't like when tools advertise `${thing} at the speed of though` because I don't really think that fast.

PSA. If `pnpm/action-setup` is failing in your Github actions you need to upgrade it to `pnpm/action-setup@v4`.

I wrote a little thread🧵 with the current i18n SEO best practices .

I really like the idea of Chrome's `window​.ai` api. It enables frontends to call models installed locally on the users's device. However I am a bit worried about locking in an interface for how we call models. This is still an active area of experimentation.

RT @Rich_Harris: happening in ~25 mins 👇 come by, bring questions etc.

Really torn between static (json) & dynamic (js/ts) configs. - Static configs are way easier to edit programmatically.- Dynamic configs are way better for defining plugins.

RT @nathanbrachotte: ParaglideJS is an excellent i18n solution and I'm surprised I've never seen anyone mention it here.