Letting you chose between:
- explain
- fix
- randomize

In theory it is not impossible to keep picking "randomize" as a way to complete your work.

So many where I was sure they were real :D

I guess following the specs and bike shedding very closely kinda poisons the well😅

CSS or BS? Think you know CSS? Real property or made-up nonsense? 20 rounds. No mercy.

Rather than the same old boring internet pranks, I thought I'd build something more fun this April Fools.

CSS or BS. Can you tell your CSS properties names from BS?

www.keithcirkel.co.uk/css-or-bs

This is also why the Stylelint issue is leaning towards a fix where resolving and loading other stylesheets isn't done in Stylelint itself. But rather a fix where users have an API to pass already parsed stylesheets as context.

The largest issue is how bundlers treat duplicate imports. Most default to a model that is more similar to ES modules where the first import applies. While native CSS is different, where each import is applied and so effectively the last one has effect.

This problem was also part of why I started checking how various CSS bundlers work. In theory they can all follow the CSS spec for at-import.

If all bundlers work the same way, then other tools (like linters) can apply better heuristics.

github.com/romainmenke/...

See: github.com/stylelint/st...

Very cool! Thank you for sharing ✨

www.theaddresses.com/book-now/#av...

First need to select a location, then the date picker appears. There is also a native date picker on the side 🤷

Recently got a bug report from a client who decreased zoom to 80% or 90%. Turned out to be a Chrome bug: issues.chromium.org/issues/47776...

This is something I never really considered and don't test for. Only test 100%-200% zoom per page/site and increased font size settings in browser.

I can imagine that this also becomes slow very quickly if elements have complex/deep dependency graphs.

We don't have that.

No, custom VSCode extensions, we do all our dev feedback, utilities, ... in custom extensions.

We have a single mono repo and predictable paths per project.
So at any given time the extension only needs to do a shallow scan of a single directory and only read a couple files.

I can imagine that there are different scales :)

We typically have 1-5 custom elements per project as we do server side rendering and our projects are text heavy with few interactive elements.

Doing this analysis on the fly is really fast for us.

a) no

b) yes

Why have a separate file and format when static analysis could be done one the element code itself.

Kinda looks like combinatorial explosion 🤔
Weird!

Are there sometimes sourcemaps in such cases?

haha, is there a clue as to why they create such selectors?

- layers polyfill?
- CSS nesting got out of hand?
- ...?

I don't think your points were fully represented on the call 😅.
When this was being discussed I was more in favor of keeping `auto`/`match-auto`.

Changed my mind after again reading through github.com/w3ctag/desig...

Also noticed this.
More and more people seem to get "sucked in".

Sunk cost fallacy and addiction/dependency is what it looks like to me.

Very happy that we have a "no-LLM's" policy at work 😅
I don't want to manage/review others who use LLM's.

@apply in style attribute ...

codepen.io/romainmenke/...

There is always a way....

postcss-preset-env 11.1.0 is out now.
- very basic support for mixins
- esm only (since v11)

github.com/csstools/pos...

preset-env.cssdb.org/playground/#...

require(esm) in Node.js: from experiment to stability More than a year ago, I set out to revive require(esm) in Node.js and landed an experimental implementation. After a lot of iteration and battle-testing, require(esm) is now unflagged across all suppo

Finished two retrospective blog posts on the journey of require(esm) before 2025 ends:

joyeecheung.github.io/blog/2025/12...

joyeecheung.github.io/blog/2025/12...

Interactive Playground - PostCSS Preset Env - CSS Tools Play with PostCSS Preset Env in your browser

Interested in how much the :not(#\#) selector appears in the wild.

preset-env.cssdb.org/playground/#...

okforrealnowlch(...)

Because oklch has been stable for too long in all engines...

Screenshot showing the network tab in Google Chrome. It shows that gtag js requires 155kb to download. The entire site itself (including gtag js) is 279kb.

Google Analytics / gtag scripts are way way WAAAAYYY too massive.

We mostly avoid it but some clients still request it.
In a recent project (large site) the home page above the fold is 279kb. 155kb of that is Google bloat....

A minimal script for those that do not use all features would be nice. 🙇

Is there a text version or link?

I agree that it is confusing. I just don't read the keywords as they are intended.

I always read "element-shared" as "specific to this element" instead of "shared between elements"

A fairly recent spec edition is the alpha() function: drafts.csswg.org/css-color-5/...

So the "function --lower-opacity" example will eventually be simpler still 🎉

[css-conditional] Element queries · Issue #10509 · w3c/csswg-drafts Container queries allow us to query a container for specific conditions, including values of custom properties through style() queries. But they do not allow us to query the element itself. In #100...

I opened a spec issue about this some time ago: github.com/w3c/csswg-dr...

Always helpful if you can add some of your use cases :)

It's pretty dynamic on our end, such bots almost always trip over one of our honey pots and get banned automatically.

same here