RESOLVED FIXED Bug 43080
History.pushState() + navigation operates on top frame when called from nested context 
https://bugs.webkit.org/show_bug.cgi?id=43080
Summary History.pushState() + navigation operates on top frame when called from neste...
Justin Schuh
Reported 2010-07-27 14:48:38 PDT
Created attachment 62749 [details] repro I've attached two files that demonstrate an iframe using pushState() and back() to force navigation of the parent frame. The scripts work automatically on Chrome, but you can observe the same behavior with manual navigation in Safari. Based on my read of the standard we should be navigating the iframe's nested browsing context, not the top frame. This might be a regression from bug 36435.
Attachments
repro (393 bytes, application/x-zip-compressed)
2010-07-27 14:48 PDT, Justin Schuh 		no flags
Details
simple fix (wip) (919 bytes, patch)
2010-07-27 16:53 PDT, Darin Fisher (:fishd, Google) 		no flags
DetailsFormatted DiffDiff
v1 patch (6.06 KB, patch)
2010-07-27 22:19 PDT, Darin Fisher (:fishd, Google) 		beidson: review+
DetailsFormatted DiffDiff
Show Obsolete (1) View All Add attachment proposed patch, testcase, etc.
Darin Fisher (:fishd, Google)
Comment 1 2010-07-27 16:53:47 PDT
Created attachment 62773 [details] simple fix (wip) Here's the fix. I still need to add a layout test, etc. I'm also thinking of changing createItemTree to return the target item as an out param so that we can avoid the recursive walk done by HistoryItem::targetItem().
Darin Fisher (:fishd, Google)
Comment 2 2010-07-27 22:19:05 PDT
Created attachment 62795 [details] v1 patch
Darin Fisher (:fishd, Google)
Comment 3 2010-07-30 13:40:11 PDT
Landed as http://trac.webkit.org/changeset/64369
WebKit Review Bot
Comment 4 2010-07-30 13:56:42 PDT
http://trac.webkit.org/changeset/64369 might have broken Qt Linux Release
Note You need to log in before you can comment on or make changes to this bug.