29510 2016-03-02 00:34:56 +0000 Containing block for sticky positioning should skip anonymous containing blocks. 2016-03-02 01:22:56 +0000 1 1 1 Unclassified CSS Positioned Layout unspecified All All NEW P2 normal --- 1 flackr dave.null dbaron public-css-bugzilla oldest_to_newest 125316 0 flackr 2016-03-02 00:34:56 +0000 https://www.w3.org/TR/css3-positioning/#sticky-pos states that the sticky-constraint rectangle is computed based on the containing block of the sticky positioned element, however, since a sticky position element can be inline the containing block can be an anonymous block. I believe constraining to the anonymous block would be unintuitive to a site author as it doesn't correspond to an actual element on their site. Firefox and safari's sticky implementations both seem to skip anonymous containing blocks when computing the sticky constraining rectangle as can be seen on http://flackr.github.io/compositor-worker/sticky/sticky.html. I think the spec should clarify that anonymous containing blocks are skipped. 125317 1 dbaron 2016-03-02 01:22:56 +0000 Should such anonymous boxes count as containing blocks at all? It's not clear to me that they should; it might be worth testing the other things that should depend on whether they count as containing blocks. (The spec specifically excludes percentages in the last paragraph of https://drafts.csswg.org/css2/visuren.html#anonymous-block-level , but it's not clear to me why it excludes only those and not other things.)