Re: [csswg-drafts] [css-text-decor-4] Allow to interpolate between `auto` and length values in `text-decoration-inset` (#13036)

The CSS Working Group just discussed ``[css-text-decor-4] Allow to interpolate between `auto` and length values in `text-decoration-inset` ``, and agreed to the following:

* `RESOLVED: Go with option 2 (computes as-is, but animates as length when interpolating with a length).`

<details><summary>The full IRC log of that discussion</summary>
&lt;fantasai> jfkthame: The issue is basically whether we should allow interpolation between auto value and explicit lengths.<br>
&lt;fantasai> jfkthame: Opinions were that we should, but then questions about exactly how that's computed.<br>
&lt;Rossen> 1. auto computes to some length<br>
&lt;Rossen> 2. auto computes as-is, but animates as some length (like font-style: normal animating as oblique 0deg)<br>
&lt;Rossen> 3. auto animates as-is, but we can represent interpolations between auto and a length (like calc-size())<br>
&lt;Rossen> 4. auto can't be interpolated with lengths<br>
&lt;TabAtkins> fantasai: I think jonathan brought up an interseting point, would auto need to be context dependent?<br>
&lt;TabAtkins> fantasai: like would it trim differently depending on what content is adjacent, whther it's start or end of line...<br>
&lt;TabAtkins> fantasai: with that behavior, to make it animatable we'd need this to be a two-value property, an auto flag and a length<br>
&lt;Rossen> ack fantasai<br>
&lt;TabAtkins> fantasai: but that doesn't mean we can't resolve auto to a length for th epurpose of animating<br>
&lt;TabAtkins> fantasai: so i think the first option is simplest and we shoudl go with<br>
&lt;flackr> +1<br>
&lt;TabAtkins> fantasai: if we do end up needing context-dependent value, extend the grammar to combine auto with length<br>
&lt;TabAtkins> fantasai: auto by itself would invoke a ua-dependent length<br>
&lt;TabAtkins> fantasai: so until we get to that point i think it's okay to compute to a length<br>
&lt;TabAtkins> fantasai: but if we want to leave open that eventuality... might want gCS() to still return "auto". don't know if it'll be important enough of a compat concern to care<br>
&lt;TabAtkins> jfkthame: that sounds more like Oriol's second option, where it retains as "auto" but is treated as a length in animation<br>
&lt;TabAtkins> (I agree, that does sound like the behavior being described)<br>
&lt;TabAtkins> flackr: I heard that as an argument fo rthe first option, to compute to a length<br>
&lt;TabAtkins> fantasai: i think jfkthame is right, if we want auto to ever ahve more behavior than "just a length", for future compat we'd need it to reflect as "auto" in computed style and let it turn into a length in animations<br>
&lt;TabAtkins> fantasai: but we could also just go with first option for now and change it in the future if we need<br>
&lt;TabAtkins> Rossen: what would gCS() return?<br>
&lt;TabAtkins> fantasai: just a length<br>
&lt;TabAtkins> Rossen: wouldn't that be hard to chagn elater?<br>
&lt;TabAtkins> flackr: in theory, but it's unlikely to be a practical compat issue<br>
&lt;TabAtkins> (I like going with option 1 for simplicity for now)<br>
&lt;fantasai> same<br>
&lt;fantasai> Rossen: 2nd is more conservative option<br>
&lt;fantasai> Rossen: But if we're not concerned about the impact here, then either should work<br>
&lt;fantasai> Rossen: 2 is path to 1 in my mind<br>
&lt;fantasai> Rossen: So do we start with option 2 or just go for option 1<br>
&lt;fantasai> jfkthame: I would lean towards 2. Having 'auto' appear in gCS() seems more expected.<br>
&lt;fantasai> jfkthame: One thing that's surprising about it is that if we have 'auto' in gCS() but then we animate it as a length, then if you animate from auto to auto you'll see the value as a length, which could be surprising<br>
&lt;fantasai> jfkthame: but I don't think it matters really<br>
&lt;fantasai> Rossen: objections to option 2?<br>
&lt;flackr> we can say that interpolation of auto to auto is auto<br>
&lt;fantasai> PROPOSED: Go with option 2 (computes as-is, but animates as length when interpolating with a length).<br>
&lt;fantasai> RESOLVED: Go with option 2 (computes as-is, but animates as length when interpolating with a length).<br>
</details>


-- 
GitHub Notification of comment by css-meeting-bot
Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/13036#issuecomment-4128224863 using your GitHub account


-- 
Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config

Received on Wednesday, 25 March 2026 17:00:27 UTC