26584 2014-08-14 13:47:03 +0000 "fully exit fullscreen" is not invoked for normal navigation 2015-10-06 08:00:53 +0000 1 1 1 Unclassified WHATWG HTML unspecified Other other RESOLVED MOVED http://www.whatwg.org/specs/web-apps/current-work/#the-history-interface blocked on dependencies P3 normal Unsorted 26813 1 contributor ian annevk ian mike philipj roc contributor oldest_to_newest 110178 0 contributor 2014-08-14 13:47:03 +0000 Specification: http://www.whatwg.org/specs/web-apps/current-work/ Multipage: http://www.whatwg.org/C#the-history-interface Complete: http://www.whatwg.org/c#the-history-interface Referrer: Comment: "fully exit fullscreen" is not invoked for normal navigation Posted from: 185.26.182.28 User agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/37.0.2062.35 Safari/537.36 OPR/24.0.1558.21 (Edition Developer) 110179 1 philipj 2014-08-14 13:49:19 +0000 "fully exit fullscreen" is only invoked by "traverse the history by a delta", which is used for go(), back() and forward(). The fullscreen spec claims that "HTML invokes the fully exit fullscreen algorithm for navigation." which isn't entirely accurate. What's the point of exiting fullscreen in these cases but not e.g. when the user followsa link in fullscreen? 110180 2 philipj 2014-08-14 13:50:16 +0000 Context/history: http://krijnhoetmer.nl/irc-logs/whatwg/20140814#l-567 https://www.w3.org/Bugs/Public/show_bug.cgi?id=18840 110181 3 philipj 2014-08-14 13:50:57 +0000 annevk and roc probably have opinions. 110202 4 roc 2014-08-14 22:26:47 +0000 I don't remember trying to distinguish different navigation methods. I thought we should exit fullscreen on all toplevel navigations. Basically I'm worried that if you go fullscreen and we show a message about domain A being fullscreen, it's easy to navigate that browsing context to domain B without A necessarily blessing that navigation. For example if you click on an ad. Also I think if A was window.opened by B, then B can later navigate that window to another B page. 110203 5 philipj 2014-08-14 22:42:04 +0000 (In reply to Robert O'Callahan (Mozilla) from comment #4) > I don't remember trying to distinguish different navigation methods. I > thought we should exit fullscreen on all toplevel navigations. Is that top-level only? Should navigating an iframe (fully) exit fullscreen? I don't think the HTML spec makes the distinction, but I'm not very familiar with navigation. How about when a popup is opened while in fullscreen? Chromium will exit fullscreen in that case: https://code.google.com/p/chromium/issues/detail?id=304547 > Basically I'm worried that if you go fullscreen and we show a message about > domain A being fullscreen, it's easy to navigate that browsing context to > domain B without A necessarily blessing that navigation. For example if you > click on an ad. Also I think if A was window.opened by B, then B can later > navigate that window to another B page. Do you (want to) distinguish between same-origin and cross-origin navigation? 110204 6 roc 2014-08-14 23:54:01 +0000 (In reply to Philip Jägenstedt from comment #5) > (In reply to Robert O'Callahan (Mozilla) from comment #4) > > I don't remember trying to distinguish different navigation methods. I > > thought we should exit fullscreen on all toplevel navigations. > > Is that top-level only? Should navigating an iframe (fully) exit fullscreen? I don't think navigating an iframe should exit fullscreen, because the domain the user is trusting has not changed. > How about when a popup is opened while in fullscreen? Chromium will exit > fullscreen in that case: > https://code.google.com/p/chromium/issues/detail?id=304547 No opinion on that. > > Basically I'm worried that if you go fullscreen and we show a message about > > domain A being fullscreen, it's easy to navigate that browsing context to > > domain B without A necessarily blessing that navigation. For example if you > > click on an ad. Also I think if A was window.opened by B, then B can later > > navigate that window to another B page. > > Do you (want to) distinguish between same-origin and cross-origin navigation? No opinion on that either. 110207 7 philipj 2014-08-15 08:14:39 +0000 It sounds like at a minimum, cross-origin top-level navigation should fully exit fullscreen, regardless of how that navigation happens. That doesn't match the spec. 110260 8 ian 2014-08-18 20:50:01 +0000 Weird. Why is this in "traverse history by a delta" and not just "traverse"? 110263 9 philipj 2014-08-18 21:18:14 +0000 Weird indeed. It's been like that since the beginning: https://www.w3.org/Bugs/Public/show_bug.cgi?id=18840 http://html5.org/tools/web-apps-tracker?from=7563&to=7564 If you don't know why, I don't think anybody does :) 111561 10 ian 2014-09-15 23:25:08 +0000 Added XXX26584 markers to the spec while I wait for bug 26813. Basically I just need to add something to "update the session history with the new page" and to "discard a Document", I think, that is similar to what's already in the spec. Either that, or move the text from the current place to the "traverse" algorithm and the "discard" algorithm, not quite sure yet. 111563 11 contributor 2014-09-15 23:32:24 +0000 Checked in as WHATWG revision r8774. Check-in comment: Add some internal editorial notes for forthcoming fullscreen edits https://html5.org/tools/web-apps-tracker?from=8773&to=8774 123495 12 annevk 2015-10-06 08:00:53 +0000 https://github.com/whatwg/html/pull/230