- From: CSS Meeting Bot via GitHub <noreply@w3.org>
- Date: Thu, 21 Aug 2025 14:55:32 +0000
- To: public-css-archive@w3.org
The Houdini Task Force just discussed `[css-anchor-position] Add containing block rules to acceptable elements in top layer`, and agreed to the following: * `RESOLVED: top layer rules is just about DOM order restriction and top level viewport is above top layer CBs` <details><summary>The full IRC log of that discussion</summary> <alisonmaher> @dshin: in find in anchor algo, one condition of valid anchor, if you refer to one in top layer, because elements in top layer are abspos and fixedpos, they define their on CB<br> <alisonmaher> ...?? even if anchor in ?? we don't know pos because static pos above it<br> <alisonmaher> ...top layer shouldn't factor into if it is valid anchor<br> <alisonmaher> emilio: things in top layer don't have static pos<br> <alisonmaher> TabAtkins: this is about what anchors are avail<br> <alisonmaher> ...it looks like browsers if you have an abspos of a child of top layer, the top layer does diff than initial CB<br> <alisonmaher> emilio: they establish CB for abpos because top layer elements need to be abspos<br> <alisonmaher> TabAtkins: right<br> <alisonmaher> ...if you are an abspos in top layer, your CB is trapped in top layer so you can't anchor outside<br> <alisonmaher> ...not case for fixedpos since root scroller is outside<br> <alisonmaher> emilio: top layer is a sibling after root scroller<br> <alisonmaher> TabAtkins: proposed resolution in the issue<br> <alisonmaher> @dshin: makes sense as long as top layer isn't mentioned<br> <alisonmaher> TabAtkins: doesn't work because of fixedpos ?? a notion of DOM order that generalizes to top layer needs to happen<br> <alisonmaher> emilio: should use box order but we don't have it defined for abspos<br> <alisonmaher> ...the big change is that regular abspos works just by virtue of top layer requiring abspos so no special case<br> <alisonmaher> PROPOSED: top layer rules is just about DOM order restriction and top level viewport is above top layer CBs<br> <alisonmaher> emilio: should be layout order but fine. If you have two top layer items ?? should be concept for this<br> <alisonmaher> RESOLVED: top layer rules is just about DOM order restriction and top level viewport is above top layer CBs<br> <alisonmaher> fantasai: related to this, someone posted a test case where 2 abspos and anchor between them<br> <alisonmaher> ...they noticed only worked if anchor pos box is third box<br> <alisonmaher> TabAtkins: correct<br> <alisonmaher> fantasai: possible to relax to pos to later if the box is not anchor positioned<br> <alisonmaher> TabAtkins: no because the painting order and sizing order of abspos is well defined and compat sensitive<br> <alisonmaher> ...layout in order of CB ancestors<br> <alisonmaher> fantasai: layout of abspos is not order dependent<br> <alisonmaher> emilio: as long as you perform static pos layout first<br> <alisonmaher> fantasai: as long as not nesting abspos or anchoring, order doesn't matter<br> <alisonmaher> ...would it be possible to allow thi<br> <alisonmaher> TabAtkins: would need to update signal<br> <alisonmaher> ...things that aren't achored would be put before those that aren't<br> <alisonmaher> ...but in middle of layout not style at that point<br> <alisonmaher> ...some people will find reordering DOM annoying<br> <alisonmaher> TabAtkins: just reorder DOM<br> <alisonmaher> emilio: anchor vs non anchored doesn't work with nesting<br> <alisonmaher> ...that is non trivial check<br> <alisonmaher> ...prefer to keep tree order<br> <alisonmaher> ...that is the order the boxes are in<br> <alisonmaher> ...no extra sorting<br> <alisonmaher> fantasai: seems likely to be a complaint<br> <alisonmaher> emilio: we can revisit. Might be possible but tree order if easy to understand<br> <alisonmaher> ...it is also the painting order, weird to diverge from layout order<br> <alisonmaher> TabAtkins: don't like anchor pos to anchor to lower abspos and abpos re-positions and layout breaks. We shouldn't break<br> <alisonmaher> ...state of anchor or not is a valid category for I am a valid anchor<br> <alisonmaher> ...need dependency tracking to avoid cycling or do tree order<br> <TabAtkins> and I don't want to do dependency tracking :)<br> </details> -- GitHub Notification of comment by css-meeting-bot Please view or discuss this issue at https://github.com/w3c/csswg-drafts/issues/11602#issuecomment-3210951128 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Thursday, 21 August 2025 14:55:33 UTC