15294 2011-12-21 08:37:40 +0000 [Shadow]: Tree composition algorithm should select the child nodes of the shadow host at a shadow insertion point when there is no older shadow tree 2011-12-21 19:32:50 +0000 1 1 1 Unclassified WebAppsWG HISTORICAL - Component Model unspecified PC All RESOLVED FIXED P2 normal --- 14978 15300 1 dominicc dglazkov public-webapps-bugzilla oldest_to_newest 61861 0 dominicc 2011-12-21 08:37:40 +0000 The tree composition algorithm step 5.2.1.1 says "If there is no older tree, stop." I think that this should be changed to: "Distribute the remaining nodes in POOL, if any, to POINT." As I understand the intent of the shadow element it is to compose with other shadow trees. However since the presence of an older shadow tree is not detectable, it is not possible to create a shadow tree that is generally composable. It is Sophie’s choice, do I <shadow></shadow> — compose with other shadow trees, and let unused light children die if there is no older shadow tree or <content></content> — render unused light children, but let any older shadow trees die? If the tree composition algorithm is changed in the proposed way then the shadow element can be used to create generally composable shadow trees. Alternatively, the spec could state that elements should have an oldest shadow equivalent to <content></content>. (Should and not must for UAs which want to use shadow DOM as an implementation mechanism, for example, to suppress the rendering of children in elements which do not permit child content etc.) 61887 1 dglazkov 2011-12-21 19:27:55 +0000 (In reply to comment #0) > The tree composition algorithm step 5.2.1.1 says "If there is no older tree, > stop." > > I think that this should be changed to: > > "Distribute the remaining nodes in POOL, if any, to POINT." I think this is what I'll do for now. > Alternatively, the spec could state that elements should have an oldest shadow > equivalent to <content></content>. (Should and not must for UAs which want to > use shadow DOM as an implementation mechanism, for example, to suppress the > rendering of children in elements which do not permit child content etc.) I really like this notion, but it has exactly the same effect as the first change you proposed. Perhaps we'll come back to doing it this way later. 61888 2 dglazkov 2011-12-21 19:30:49 +0000 http://dvcs.w3.org/hg/webcomponents/rev/2c3c3971a9f5