[css-anchor-position-1] Should custom property declarations be valid in @try?

[cdoublev]

I initially asked this in #9150.

@xiaochengh warned that accepting them would break anchor positioning entirely:

.target {
  position: var(--x);
  --x: absolute;
  position-fallback: --fallbacks;
}

@position-fallback --fallbacks {
  @try { --x: static; }
}

But I wonder if it would work if custom properties are accepted only if there is one or more declarations of the other accepted properties?

@position-fallback --fallbacks {
  @try { --x: static; } /* invalid, ignored */
  @try { --x: static; width: 100px } /* valid */
}

I do not have a strong opinion and I am only interested in disallowing custom properties, because @try is the only rule accepting property declarations that do not accept custom properties.

[xiaochengh]

But I wonder if it would work if custom properties are accepted only if there is one or more declarations of the other accepted properties?

I don't see how this makes a difference. It still allows custom properties to set arbitrary other properties.

[cdoublev]

It still allows custom properties to set arbitrary other properties.

Right, of course, you want fallbacks that can only change the position of the anchored element. You want to know if a fallback would make it overflow its containing block without re-computing styles of all elements.

[cdoublev]

I hope it is ok if I re-open it, because custom properties are implicitly accepted in keyframe rules, so I think either CSS Anchor Position or CSS Animation should explicitly exclude or include them.

---

From this IRC log:

custom property use in @Try is a good consideration

But this does not seem to have been discussed further. Unless otherwise stated, I am goind to assume this is still not a good consideration.

[andruud]

The spec excludes them by not including them in the list of things that are accepted. That seems good enough?

[cdoublev]

Yes, my thinking was wrong. Sorry. Definitely closed. 🔒