- From: Boris Chiou via GitHub <noreply@w3.org>
- Date: Sat, 09 May 2026 08:44:09 +0000
- To: public-css-archive@w3.org
OK, I found the answer of my first question. > The spec only says we have to find keyframe offset 0% and 100% (instead of computedOffset), so I am confused about this. Should we have to generate the implicit frames if we can find a keyframe (with range names) whose computedOffset is 0% or 100%? So per scroll-animations-1, the [section for keyframe selector](https://drafts.csswg.org/scroll-animations-1/#named-range-keyframes): > It is possible that these attachment points are outside the [active interval](https://drafts.csswg.org/web-animations-1/#active-interval) of the animation; in these cases the automatic from (0%) and [to](https://drafts.csswg.org/css-shapes-1/#valdef-shape-to) (100%) keyframes are only generated for properties that don’t have keyframes at or earlier than 0% or at or after 100% (respectively). We don't have to generate the initial/final keyframes if the attachment points are just equal to 0% or 100% as well. So in my previous example, I think it is WebKit's bug. We shouldn't generate the 100% keyframes because there is a keyframe whose attachment point is at 100%. About the second question, I'd like to just follow Blink and WebKit for Gecko: All the keyframes with `<timeline-range-name>` are after the generated final keyframe in the computed order. (i.e. ignore this sentence "other than a generated final keyframe" since no browser follows it now.) -- GitHub Notification of comment by BorisChiou Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/13872#issuecomment-4412070070 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Saturday, 9 May 2026 08:44:13 UTC