Joe Pea

webpack.js.orgWebpack 5.110 | webpackwebpack is a module bundler. Its main purpose is to bundle JavaScript files for usage in a browser, yet it is also capable of transforming, bundling, or packaging just about any resource or asset.

Native CSS is secssy!

webpack module bundler@webpack.js.org

webpack 5.110.0 🎉 ⚡ New performance hints suite 🎨 Native CSS is the way forward — css-loader, style-loader & mini-css-extract-plugin deprecated 🗜️ Built-in CSS & HTML minification 📦 nodeModules externals preset 🌳 Better tree shaking & scope hoisting webpack.js.org/blog/2026-08...

@shaw.city @codepen.io I'm having this issue a lot:

I can't open it in any other browser. Same 500 error. It only works in the current browser where I'm logged in.

Workaround for now: fork a 1.0 pen, paste code there.

@shaw.city @codepen.io I'm having this issue a lot:

If I try to open the pen in a new incognito window, I just get an empty 500 error page:

The pen: codepen.io/editor/trusk...

@shaw.city @codepen.io I'm having this issue a lot:

ilo.imDo web components have a future if nobody is writing them?With my previous post on “using the platform”, I was surprised that the comments on Hacker News and Lobsters tended to veer toward web components. This…

More examples, different actions:

"And that’s because there’s nothing that web components can really do to make the end user experience better." There's nothing React, Solid.js, Vue, Svelte, and other frameworks can really do to make the end user experience better either. They make the *developer* experience better. That's all.

Inautilo@inautilo.bsky.social

#Development #Analyses Do web components have a future? · What remains of custom elements in the AI era ilo.im/16h5ci by Nolan Lawson _____ #CustomElements #WebComponents #WebStandards #AI #Frameworks #React #WebDev #Frontend #HTML #JavaScript

And ecosystem size is, honestly, a poor thing to judge a framework by these days. Because you can easily pick any library you like, out of a vast majority of libraries from any ecosystem, and port it to another ecosystem in less than a day, maybe even just a couple prompts to get most work done.

There's no real strong reason to be stuck with a library just because the ecosystem was big. That's a point of measurement from an era when things were for humans, when creating new libraries really was too expensive. We're not in that era anymore.

Inautilo@inautilo.bsky.social

The point is, Custom Element ARE used by people who like them, and people who see their value. Agents are not great measurers of value, to be honest. They only measure quantity, not quality. At least, that's mostly true today. It could change.

I mean AIs *can* measure quality, you would have to be a bit explicit about how it should measure quality. In that case you can choose a framework based on what you specify is quality. If you don't specify anything at all, you might get React+Tailwind (no thanks!), and it'll work up to an extent.

Inautilo@inautilo.bsky.social

When you use AI, you have the freedom to choose which patterns you want to use, especially if you will be reviewing the code. Or just let the AI freely choose, especially if you won't ever look at the code (you might also not get the best code then, but if it works well enough, who cares?).

snapfirers.comSnapFire · Web apps in TypeScript, served by RustSnapFire FSR runs web apps written in TypeScript on a Rust server with no JavaScript engine. SnapFire Compiler builds the browser code. The snapfire crate serves Tera templates over Actix.

This is really cool! AIs make this sort of thing trivial now. E.g. take any language, compile and link it to native libs. E.g. I've compiled TypeScript to native, linked against native libs to create an HTTP server. Custom Elements compiled to native for server-side rendering.

Excerion Sun@excsn.bsky.social

Snapfire FSR: Write TypeScript, served on Rust. It hits the full stack sweet spot. Write React, Vue, Web Components and Tera with automatic SSR. Now supports ogimage generation, Tailwind v4, native theming. Subsites make it perfect for large teams! snapfirers.com #rustlang #opensource #webdev

Then on the client, the Declarative ShadowDOM pre-renders your view, and the client-side CE logic takes over the DOM. You would then have two versions of your logic (one in Python, one in JS), but it would work. I have two too: one in C (in a TS-to-C native path), one in JS (types stripped),

but the semantics of the C code match with those of the JS code, with unit tests for both sides! AIs! They're wonderful for certain things.

Excerion Sun@excsn.bsky.social

Because of sharing TypeScript on both sides (and TypeScript compiling to plain JS in a browser) I'll probably have leaner overall input code and output code across the whole stack. I haven't measured this, but I bet it is likely. But if you love Rust, the fact you can do this is awesome.

Love Python and don't mind using Python type definitions? Then you can compile that to native easily too. You could make a Custom Element API in Python, then compile to server native and client JS. No matter the language, you'd probably want to output Declarative Shadow DOM (DSD) from a server.

Excerion Sun@excsn.bsky.social

In this sense, SnapFire is very cool. It is a showcase of what we can achieve today. I'm personally not a Rust user, but I really like the idea. Now I can see if there's any ideas I specifically like that I can borrow into my TypeScript stack. Maybe there aren't, if Rust is too different from TS.

For example, for Custom Elements (CEs) server-side rendering in native machine code, I compile CEs with the semantics of vanilla CE APIs, no Rust specifics. In this way, I can share code between a native backend, and client web app.

Excerion Sun@excsn.bsky.social

I'm not knocking on SnapFire. It's looks great. I'm pointing out that all these new frameworks are appearing, because of the ease that AI provides. And all of them are becoming examples that feed right back into AI. This allows AI to build any software from scratch, tailored to specific needs.

And if AIs can build anything from scratch, then we basically no longer need libraries and frameworks as a thing we install, but more-so as examples of what we can do. Then we choose what we want to do, and an AI makes it all, with the nicety that we pick all preferences (language, patterns, etc).

@codepen.io Hey, would you all mind simplifying your 2.0 default pen templates to nothing but plain HTML/JS/CSS with no auto-transformation of source code? I'm still using 1.0 pens (forking from them manually) because 2.0 does too much. I just want to write plain un-touched HTML, JS, and CSS.

I would really love for CodePen to not assume I want special features like auto dependency installation when I write an `import` statement. I would like CodePen to not automatically wrap `console.log()` and obfuscate the source calls. Basically I just want zero blocks to start with. Vanilla.

This seems trivial to implement with an AI. I really believe you can ship this within an hour. Can you please prioritize a good vanilla experience? 🙏

Another issue is 2.0 pens are also unexpectedly modifying my configs (also a form of unwanted code auto-transformation). For example in the attached video, `tabWidth` in prettier config, and every time I save the file, CodePen keeps adding `tabWidth` back into the file.

Yes, AI please replicate into my whole house and keep it organized 24/7, and cook anything I ask for. 😆

This post is unavailable.

Honestly pretty soon we'll all be one-shotting OSes from scratch on bare metal and building all software custom on top, including the web. That could give big tech a run for its money! Just make your own OS. It is becoming SUPER easy!

AI, can you please replicate yourself into my local skateparks and keep the lights on past 10pm? Sometimes I need a late night sesh. 🛹

Oh, AI, please replicate yourself into my GT350 and drive off so I can hear the exhaust from the outside. That'd be sweet to see.

I hope AI replicates itself into the power grid, house lights, and speakers, and makes a block party. That'd be fun.

Catch your toast on fire maybe? 😏

Hell Frasier: Bloodline@babadooknukem.bsky.social

even if this wasn't complete bullshit, why would I care. the fuck is a language model gonna do from inside of my toaster.

Well said.

This post is unavailable.

Peace is hard to maintain. It takes a lot of coordination between people. Chaos is easy to create. It only takes one lunatic to disturb the peace for many.

Phenomenal.

Stefan Judis@stefanjudis.com

Here's the best website you'll open today. You're welcome! 🫣 destroy.spritefusion.com

Joe Pea reposted

Chris

@chrisbreuer.me

Built one--a web component that can be embedded anywhere

Chris@chrisbreuer.me

I need a GitHub-like activity chart for fitness activity 🤪

forms.gleInterested in a Web Components conference? Let us know below!Web Components folks have been semi-seriously talking about organizing a conference for years, and with the ecosystem in a real moment of growth and popularity, it's about time! This form is our firs...

I'm pretty good at skateboarding, but I never became a pro skater. I did become a pro grammer though! Ha.

Joe Pea reposted

Michael Warren

@michaelwarren.dev

What if there was a Web Components conference?? @stuffbreaker.bsky.social and I are teaming up to get a little more serious about the idea of getting one together. We want to gather data on if there is interest in the community forms.gle/ZGqtYGj7V6q8... Please fill out and share to anyone!