23793 2013-11-11 10:24:26 +0000 "For each element e in elementSet, append the el..." 2014-02-12 21:08:10 +0000 1 1 1 Unclassified WebAppsWG DOM unspecified PC Linux RESOLVED MOVED P2 normal --- 1 simon.sapin annevk bzbarsky jackalmage mike www-dom public-webapps-bugzilla oldest_to_newest 96122 0 simon.sapin 2013-11-11 10:24:26 +0000 http://dom.spec.whatwg.org/#interface-parentnode [[ For each element e in elementSet, append the elements that are not in result and result from matching s against e's root using scoping root e, to result. [SELECTORS] ]] This should say something about pseudo-elements. In implementations they seem to be valid, but not match anything. 96168 1 annevk 2013-11-12 06:40:09 +0000 I don't understand why we need to mention them. 96201 2 simon.sapin 2013-11-12 21:30:22 +0000 Well, based on the current spec, I don’t know what is supposed to happen when I use document.query('p::after'). Does it throw? Return null? Ignore ::after and return a <p> element? It’s either specified in a part of the spec that I missed, or it’s not defined which is bad for interop. 96229 3 annevk 2013-11-13 08:52:48 +0000 How would what it is matched be ever up to DOM? That's up to the Selectors specification. 96512 4 jackalmage 2013-11-19 16:02:07 +0000 It's not a matter of what gets matched (Selectors does indeed define that), but what gets returned by the function. When you match a selector against a tree, you might get pseudo-elements. In that case, what gets added to _result_ and ultimately returned to the user? (The right answer is to use the new PseudoElement interface, I think, but the match-current-implementation behavior is to ignore them.) 96513 5 bzbarsky 2013-11-19 16:04:33 +0000 > The right answer is to use the new PseudoElement interface, I think Defining the right behavior there could get ... pretty exciting. I think the right thing here is in fact to match current implementations. 96514 6 annevk 2013-11-19 16:30:52 +0000 Wait, Selectors matches against a DOM, not a layout tree Tab. This again shows you not defining the actual tree creates problems. 96515 7 jackalmage 2013-11-19 16:44:40 +0000 (In reply to Anne from comment #6) > Wait, Selectors matches against a DOM, not a layout tree Tab. This again > shows you not defining the actual tree creates problems. Dunno what you're talking about, bro. Selectors totally returns pseudo-elements when you specify them, at least in CSS. Agree that there's still gaps there that need to be filled in, though. 96516 8 annevk 2013-11-19 17:02:12 +0000 Bro? What I talk about is that pseudo-elements are not in the DOM so there's no way to return them. 100432 9 annevk 2014-02-12 20:00:05 +0000 Is this solved with the updates to Selectors? 100437 10 simon.sapin 2014-02-12 21:08:10 +0000 It’s not clear to me that the issue is solved, but I guess DOM now says what it can/should and the issue is now in Selectors. Moved to http://lists.w3.org/Archives/Public/www-style/2014Feb/0447.html