Content API changes to improve DOM stitching in ThreatDetails code.

content/public/browser/render_frame_host.h

Index: content/public/browser/render_frame_host.h
diff --git a/content/public/browser/render_frame_host.h b/content/public/browser/render_frame_host.h
index fee89ccb9af5fd8827dbf3c5c97fcea784213afa..c0bbedd354173df75ceecbe027b9d1983961d679 100644
--- a/content/public/browser/render_frame_host.h
+++ b/content/public/browser/render_frame_host.h
@@ -63,6 +63,8 @@ class CONTENT_EXPORT RenderFrameHost : public IPC::Listener,
   // Returns a RenderFrameHost given its accessibility tree ID.
   static RenderFrameHost* FromAXTreeID(int ax_tree_id);
 
+  static int LookupOtherFrameTreeNodeId(int process_id, int routing_id);

[dmazzoni, 2017/04/21 15:48:48]

What does "Other" mean in this context?

I think a more natural API would be something like:

static RenderFrameHost* FromProcessIdAndRoutingId(int process_id, int
routing_id)

[lpz, 2017/05/02 19:36:24]

Renamed this function to be more generic.

For "other", I just used reused the wording from the safe browsing code
(other_frame_routing_id), where other refers to the fact that the contents of an
iframe are in some "other" renderer from the one that sees the iframe element.

As for returning a RenderFrameHost*, how would we support the case where the
{pid,rid} pair points to a RenderFrameProxyHost?

[dmazzoni, 2017/05/05 17:32:28]

Every RenderFrameProxyHost has an associated RenderFrameHost
for the process that actually hosts that frame, so you could
use that as the canonical one.

[lpz, 2017/05/10 14:21:08]

Done.

+
   ~RenderFrameHost() override {}
 
   // Returns the route id for this frame.