Today, frame's id is its address. Its propagation to the javascript is fishy (reinterpret cast + assumption that it fits). We should either introduce a counter for frames or use string identifier.
Created attachment 86258 [details] Patch
Created attachment 86259 [details] [PATCH] With right errors removed from header.
Attachment 86258 [details] did not build on chromium: Build output: http://queues.webkit.org/results/8200683
Attachment 86258 [details] did not build on mac: Build output: http://queues.webkit.org/results/8207277
Comment on attachment 86259 [details] [PATCH] With right errors removed from header. lgtm
Comment on attachment 86259 [details] [PATCH] With right errors removed from header. View in context: https://bugs.webkit.org/attachment.cgi?id=86259&action=review > Source/WebCore/inspector/InspectorResourceAgent.cpp:236 > + unsigned long long address = reinterpret_cast<uintptr_t>(pointer); does unsigned long long always fit pointer type? > Source/WebCore/inspector/InspectorResourceAgent.cpp:238 > + return String::format("%.0llX", address); We should use a map address->id to enumerate all iframes instead of exposing raw address over the protocol.
Comment on attachment 86259 [details] [PATCH] With right errors removed from header. Clearing flags on attachment: 86259 Committed r81571: <http://trac.webkit.org/changeset/81571>
All reviewed patches have been landed. Closing bug.