Set RemoteDOMWindow's initial opener.
Created attachment 338160 [details] WIP Patch
As I mentioned to Chris elsewhere, I think I might consider having RemoteDOMWindow::opener() return a WindowProxyController (and perhaps rename it to WindowProxy or DOMWindowProxy) since a WindowProxyController is just the window proxy part of an AbstractFrame. I would further change everything (in the IDLs) that returns a DOMWindow to return a WindowProxy instead.
(In reply to Sam Weinig from comment #2) > As I mentioned to Chris elsewhere, I think I might consider having > RemoteDOMWindow::opener() return a WindowProxyController (and perhaps rename > it to WindowProxy or DOMWindowProxy) since a WindowProxyController is just > the window proxy part of an AbstractFrame. I would further change > everything (in the IDLs) that returns a DOMWindow to return a WindowProxy > instead. I think I could return a WindowProxyController instead of an AbstractFrame in this patch, I like the idea and this is a small change. Updating all the IDLs to use WindowProxy is definitely the way to go, I agree but I think this is out of scope of this patch. I can take care of it in a follow-up.
(In reply to Sam Weinig from comment #2) > As I mentioned to Chris elsewhere, I think I might consider having > RemoteDOMWindow::opener() return a WindowProxyController (and perhaps rename > it to WindowProxy or DOMWindowProxy) since a WindowProxyController is just > the window proxy part of an AbstractFrame. I would further change > everything (in the IDLs) that returns a DOMWindow to return a WindowProxy > instead. Note that forcing the implementation to return a WindowProxyController means we have to null check the frame. Previously, we'd just return a potentially null frame and toJS() would take care of the null check for us.
Created attachment 338251 [details] Patch
Comment on attachment 338251 [details] Patch Very nice.
Comment on attachment 338251 [details] Patch Clearing flags on attachment: 338251 Committed r230789: <https://trac.webkit.org/changeset/230789>
All reviewed patches have been landed. Closing bug.
<rdar://problem/39543247>