Lets go back 2 years ago: http://trac.webkit.org/changeset/102961/trunk/Source/WebCore/inspector/InspectorInstrumentation.cpp A chromium engineer changed InspectorInstrumentation::didFinishLoadingImpl to expect monotonic time instead of wall clock time. The argument was renamed to monotonicFinishTime. The header was never changed (still says finishTime instead of monotonicFinishTime) and neither was any other port changed to pass the new expected time format. Fast forward to today. Sigh. This causes end times to essentially double because convertMonotonicTimeToDocumentTime will add the finish time (wall clock time) to the document start time (also wall clock time), minus a reference monotonic time (a smaller number than wall clock time). Most call sites pass 0 for finishTime, so this just caused currentTime to be used later down the line. Thanks Google.
<rdar://problem/15793435>
Created attachment 220856 [details] Patch
Comment on attachment 220856 [details] Patch r=me
Comment on attachment 220856 [details] Patch Clearing flags on attachment: 220856 Committed r161670: <http://trac.webkit.org/changeset/161670>
All reviewed patches have been landed. Closing bug.