Does this need an intent to implement?

We shouldn't write an intent to implement more -webkit- stuff, and we really shouldn't add any more -webkit- stuff to the platform.

Better to live with the FIELDSET / LEGEND magic until we have a standardized way of disabling it.

Why not?

https://github.com/whatwg/html/pull/3934 is the standardized way of disabling it.

That said, fixing the spec for appearance and unprefixing is tracked in https://github.com/w3c/csswg-drafts/issues/3024

https://www.chromium.org/blink#vendor-prefixes

I say we wait for https://github.com/w3c/csswg-drafts/issues/3024 to resolve. A good start might be to add what's currently supported with -webkit-appearance (but not add new stuff) to https://compat.spec.whatwg.org/ , right?

OK, though -webkit-appearance is now cross-browser, so it's not a new property we're exposing. Only supporting the new values behind a flag sounds reasonable.

Morten, I don't think it's reasonable to forever maintain two sets of appearance values, those that can be used with -webkit-appearance, and those that can be used with an unprefixed variant. It seems better to keep those properties identical and therefore add them to -webkit-appearance and unprefix the whole property as well. (Note that I don't think we can ever remove -webkit-appearance.) Or did you have something else in mind?

Yes, the web will probably be stuck with -webkit-appearance forever. That's why we have https://compat.spec.whatwg.org/ to document such things, right? Having one "-webkit-appearance" and one "appearance" that aren't aliases definitely sounds bad, and I really hope we can avoid that. If we unprefix the property and add the new values for fieldset and legend, we should obviously allow them for -webkit-appearance as well.

My point is that we should unprefix first, then add the new values.

Would you be ok with adding the new values behind the experimental web platform features flag without waiting for unprefixing to happen?

If that's useful for anyone, that sounds fine. Can you please state the motivation for this (that's a required part of an intent to implement)?

Not being able to reset styles on fieldset/legend so they act like normal elements is a major pain point for web developers, which results in them using divs instead, and that is bad for accessibility. 

https://twitter.com/fvsch/status/1022371862674911232
https://twitter.com/etportis/status/1022250278685356032

appearance: none (or -webkit-appearance: none) is what most would reach for to disable the magic of fieldset and legend.

https://github.com/whatwg/html/issues/3912

Mozilla has a wip patch at https://bugzilla.mozilla.org/show_bug.cgi?id=1483787

No public signals for Safari/Edge, but they are aware of this proposal.

Thank you!

It would be reassuring to have some sort of spec for 'appearance' before filing an intent. It'd also be great to know what Safari and Edge think about this.

Also: https://compat.spec.whatwg.org/ should mention -webkit-appearance.

Made a PR for compat, at https://github.com/whatwg/compat/pull/104

Spec and tests for -webkit-appearance landed.

Taken out from the spec:
https://github.com/whatwg/html/pull/3934/commits/7c24f1fb60ac92d3fac1bad2ab5e3bbf47817aef