Status update for Skeleton v3

[endigo9740]

Hey everyone, Chris here from the Skeleton team. As you may know we've been hard at work putting the final touches on Skeleton v3. This has been a monumental effort that's taken just over a year. We've had major contributions from various folks in and outside the team to make this happen. The result is a fresh new version of Skeleton that provides more focus and a much better development experience.

That said, we have some good and bad news in regards to our upcoming launch date. This may include a potential delay. Please read below to understand the circumstances, please use the link provided below to vote us know what path forward you would prefer.

The Good News

The good news is we are right at the end of our queue for the v3 milestone issues. In fact, we're down to around 10 items remaining, with most already in motion. As such, we were leading up to a planned reason next Tuesday (Jan. 28). We've been rolling out update to the new Skeleton CLI migration tooling, new documentation updates, and more to prepare for this.

The Bad News

The bad news is Tailwind has thrown us a bit of a curve ball in the 11th hour with the release of Tailwind v4. It's a double edge sword for us. On one hand, we're super excited about the new direction they are taking with configuration, extensibility, and all the new features. This includes OKLAB color support, container queries, and more. Unfortunately, the timing could not be worse for us. While we've been in crunch mode for the Skeleton v3 release, Tailwind went through a fairly rapid beta phase and the new major version is now live. To understand how that might impact Skeleton, let's review what's changed.

What does Tailwind v4 change?

I've provide a high level overview of the in this thread. To summarize:

• In Tailwind v3, configuration and plugins were handled with Javascript assets
• In Tailwind v4, configuration is handled almost entirely in CSS.
• The concept of plugins are gone; replaced instead with CSS stylesheets
• However, Skeleton core package is our Tailwind plugin; which enables the overwhelming majority of our features

How does this affect Skeleton v3?

Here's what features are handled by Skeleton v3's Tailwind plugin:

• All global (aka "base") styles (background colors, etc)
• All design system features (themes, colors, presets, typography, spacing)
• All Tailwind Components (buttons, cards, form styles, tables, etc)
• Plus a number of utility classes and design tokens (documented here)

In other words Skeleton doesn't exist without this plugin. It's the heart of the entire library, and nearly everything depends on this. Even components.

Why this is a Problem now?

Plain and simple, Skeleton v3 in it's current status cannot support Tailwind v4. We've had multiple members of the team testing and it is just not in a viable place right now.

Add to that, it may not be a trivial update to transition to Tailwind v4. And in fact there's a lot of features we might want to support while we're at it, for example:

• Making use of container queries for smarter responsive styling
• Converting all theme colors to utilize OKLAB instead of RGB
• Ensuring all existing CSS styles we generate with Javascript and ingest into our plugin are instead written to stylesheets that are provided to end users.
• Ensuring all existing CSS stylesheets we ingest into our plugin remain functional and have the appropriate @directives
• Removing and replacing discontinued utility classes
• Making use of all the new variants (ex: not-*)
• Etc

While we think each of these efforts individually might go quickly, the total effort is too hard to estimate upfront. Plus tooling we rely on is already moving forward - such as the Svelte sv CLI now defaulting their add-on to Tailwind v4. The only current work around for this would be to pin to a set version of sv, which means all add-on packages will slowly grow more and more out of date over time.

Considering a Delay

This is where the community comes in. The team and I are having a hard time deciding how best to move forward with this. Currently we're considering one of two options:

1. We move forward with the release of Skeleton v3 next week as planned, but utilize a pinned version of sv, and make it clear we do not yet support Tailwind v4.
2. Alternatively, we implement a delay and go through the dev cycle to rewrite our plugin to be compatible with Tailwind v4. Our best guess is this would likely be a delay of multiple weeks.

While the former would get things out to you sooner, it would mean multiple breaking migrations to deal with. First with Skeleton v3, then later with a potential Skeleton "v4" migrating to support Tailwind v4. Where as if we delay, you'll only encounter the pain of a breaking migration a single time.

How to Vote

We want to hear from you - please use the poll linked below to let us know which strategy you prefer.

https://strawpoll.com/B2ZB95EaAgJ

Current Plans

We're not sitting on our hands while we await your feedback though. The team is already hard at work trying to find an alternative way to bridge support to Tailwind v4. We are considering and testing every option we can, and if a viable option exists we will find it. But until then, let us know what you think using the poll above.

If you have questions about anything detailed above, please reach out in the comments. And while obviously no one wants a delay, we do ask that you keep things professional and respect the opinions of folks that might differ from your own. We all want the same thing - which is to see the Skeleton update asap! Please and thanks!

[Zakrok09]

Kudos for keeping the community involved and asking for our opinion. Good luck with whatever option comes next! 🚀

[endigo9740]

Absolutely, we build Skeleton for the community, so this sort of feedback is crucial for us. And this is a very critical fork in the road, but I appreciate everyone's patience while we figure this out. I'm sure the end result will be worth the wait!

[blurskye]

absolutely cooked situation

[MrVauxs]

Given that Skeleton v3 is more or less a Svelte 5 update, I'd say release it now, patch it later.
Forever chasing the newest and latest versions of multiple "async" packages isn't preferable IMO

[endigo9740]

Just to be clear, that was our original intention @MrVauxs, but this was based on a couple assumptions:

1. That Tailwind v4's launch would come later than it did. We expected it a few months out from now
2. That and the Skeleton Tailwind plugin would be at least somewhat compatible

The reality is that Tailwind's beta went by very quickly and the full release dropped this week. Pair that with the fact that sv (the Svelte CLI) is about to make Tailwind v4 the default and ONLY option available means we'll have to enforce a pinned/dated version to ensure Tailwind v3 is the only version people are installing. But this affects ALL add-ons, including Drizzle, Playwright, Vitest, etc. The longer the gap, the more out of date these become.

Plus the changes required for Tailwind v4 support are firmly in the category of "breaking", mainly due to the difference in configuration, so they will have to come as part of a major release. The result would be we could launch Skeleton v3 in it's current status next week, then in let's say 4-6 weeks we launch a Skeleton v4 with these update. Then require folks to go through yet another major release and migration. Unfortunately any migration is a major ask of a userbase, which is why we've reached out to the community to discuss the idea of a delay.

[MrVauxs]

Hm. I imagined the timeline to be much longer than a few weeks. This makes the delay a lot more palatable, given this will be one less breaking transition.

[nullpointerexceptionkek]

I have had way too many migrations recently.

In my opinion, the few core maintainers are not focusing on V2, and issues are sometimes overlooked. If Tailwind 4 doesn't make breaking changes like renaming skeleton design tokens or breaking existing components, I wouldn't see why it would be an issue not to support it initally.

[grawlich]

Since it seems like reworking the theme system to be compatible with Tailwind v4 would be a massive undertaking, I'd be fine with waiting if that's that path you'd like to take. I can't say what would be optimal for your team, but as some previously mentioned, handling more migrations after the fact doesn't seem ideal. Given that I also recently hit the launch of a new project, ready to take a break, and had to immediately start bug fixing; I understand that launching with v3 and immediately starting work on upgrading would be pretty stressful.

I guess my conclusion is: I don't mind waiting for the upgrade to Tailwind v4 in general, but I'd rather not have to install v3 and migrate to v4 later. I'll do what I must, though. The ease of use of the theme system is what I love most in this toolset!

[endigo9740]

Thanks for the concern @grawlich. You're right, delays are never fun, but right now my priority is shipping Skeleton in the best state we can. This is the entire reason the v3 update is as large as it is, we're doing our best to push everything forward. There's no doubt that this is terrible timing, but sometimes that's how it goes. We're taking it stride and I'm confident that when we do ship, it'll be worth the wait!

Also the more I review the less I think it'll be a massive overhaul. Keep an eye out for our follow up announcement tomorrow. We've got a plan that I think will help make this very digestible and ensure folks are still getting regular updates, even for the duration of the rewrite.