Fixing remaining VC++ 2015 64-bit build breaks

chrome/browser/hang_monitor/hung_plugin_action.cc

Index: chrome/browser/hang_monitor/hung_plugin_action.cc
diff --git a/chrome/browser/hang_monitor/hung_plugin_action.cc b/chrome/browser/hang_monitor/hung_plugin_action.cc
index 82142862ab0ad00c4cb6474dd4647c88ecf825b0..3029936b5d521c34a4ee72a5e6399202274335b5 100644
--- a/chrome/browser/hang_monitor/hung_plugin_action.cc
+++ b/chrome/browser/hang_monitor/hung_plugin_action.cc
@@ -123,11 +123,11 @@ bool HungPluginAction::OnHungWindowDetected(HWND hung_window,
         // exists. The property is deleted if the window becomes
         // responsive.
         continue_hang_detection = false;
-#pragma warning(disable:4311)
+#pragma warning(disable : 4311 4302)

[Nico, 2015/11/16 21:31:07]

should these warning pragmas have a comment that explains why they're safe to
do?

[brucedawson, 2015/11/17 00:24:39]

Yes. I copied over the warning from the previous file, adding the word
'truncation' to make the type of warning more explicit.

I also modified the layout of the #pragma warning on line 136 to make it
consistent.

[Nico, 2015/11/17 00:30:27]

Hm, this now says that it's a truncation, but not why that's ok. Is this snippet
only built in 32-bit builds?

[brucedawson, 2015/11/18 01:58:11]

Okay, better fix now. TetProp/GetProp take/return a HANDLE, so HandleToUint32
can be used, and some #pragmas can be removed.

There is no *actual* truncation because an int is being stored and retrieved, so
HandleToUint32 is safe.

The warnings on the reverse process (int to HANDLE) can wait for bug 554200 (I
just added a reminder).

         int child_window_message_timeout =
             reinterpret_cast<int>(GetProp(
                 hung_window, HungWindowDetector::kHungChildWindowTimeout));
-#pragma warning(default:4311)
+#pragma warning(default : 4311 4302)
         if (child_window_message_timeout) {
           child_window_message_timeout *= 2;
 #pragma warning(disable:4312)