Navigation Timing Level 2

third_party/WebKit/Source/core/frame/LocalDOMWindow.cpp

Index: third_party/WebKit/Source/core/frame/LocalDOMWindow.cpp
diff --git a/third_party/WebKit/Source/core/frame/LocalDOMWindow.cpp b/third_party/WebKit/Source/core/frame/LocalDOMWindow.cpp
index cf5a2af03e45f5ac061d31eddef52b62d989e000..a46bc007e63063d2cc8edc5cab96047865998c10 100644
--- a/third_party/WebKit/Source/core/frame/LocalDOMWindow.cpp
+++ b/third_party/WebKit/Source/core/frame/LocalDOMWindow.cpp
@@ -76,6 +76,8 @@
 #include "core/page/Page.h"
 #include "core/page/WindowFeatures.h"
 #include "core/page/scrolling/ScrollingCoordinator.h"
+#include "core/timing/DOMWindowPerformance.h"
+#include "core/timing/Performance.h"
 #include "platform/EventDispatchForbiddenScope.h"
 #include "platform/weborigin/SecurityOrigin.h"
 #include "platform/weborigin/Suborigin.h"
@@ -1402,6 +1404,10 @@ void LocalDOMWindow::dispatchLoadEvent() {
     dispatchEvent(loadEvent, document());
   }
 
+  Performance* performance = DOMWindowPerformance::performance(*this);
+  DCHECK(performance);

[Yoav Weiss, 2016/11/10 14:58:58]

FWIW, I like the DCHECK here as it helps:
a) make the assumption that performance is not null here explicit
b) debugging such cases using release builds with dcheck on, in case the
assumption somehow breaks (e.g. in new and incomplete unit tests)

[kinuko, 2016/11/10 17:55:33]

Well, I've also heard people say multiple reasons not to have redundant DCHECKs,
but this is not on our style guide and apparently different people have
different opinions-- and I'm fine with these DCHECKs too :)

[sunjian, 2016/11/11 21:48:05]

Acknowledged.

[sunjian, 2016/11/11 21:48:05]

Acknowledged.

+  performance->addNavigationTiming(frame());
+
   // For load events, send a separate load event to the enclosing frame only.
   // This is a DOM extension and is independent of bubbling/capturing rules of
   // the DOM.