Remove DumpWithoutCrashing calls for empty PageState.

content/browser/web_contents/web_contents_impl.cc

Index: content/browser/web_contents/web_contents_impl.cc
diff --git a/content/browser/web_contents/web_contents_impl.cc b/content/browser/web_contents/web_contents_impl.cc
index 5aa40a8bbfdd132dc2cd9859266fb29139cef662..1926d9d7be805d67d70497e64b5ce0c39d926270 100644
--- a/content/browser/web_contents/web_contents_impl.cc
+++ b/content/browser/web_contents/web_contents_impl.cc
@@ -11,7 +11,6 @@
 #include <vector>
 
 #include "base/command_line.h"
-#include "base/debug/dump_without_crashing.h"
 #include "base/feature_list.h"
 #include "base/i18n/character_encoding.h"
 #include "base/lazy_instance.h"
@@ -4609,12 +4608,6 @@ void WebContentsImpl::UpdateStateForFrame(RenderFrameHost* render_frame_host,
   if (page_state == frame_entry->page_state())
     return;  // Nothing to update.
 
-  if (!page_state.IsValid()) {
-    // Temporarily generate a minidump to diagnose https://crbug.com/568703.
-    base::debug::DumpWithoutCrashing();

[alexmos, 2017/02/23 18:05:12]

Ha, I actually found one of these!  See b310b2c700000000.  That's the only one I
could find though.

(The way I checked was to look up all the crashes that include
WCI::UpdateStateForFrame on the stack.  Then in the signatures, find the one for
dump without crashing.  Then look at the authors -- most of them were from Nasko
diagnosing another bug, and one was here.)

[Charlie Reis, 2017/02/28 05:30:57]

Wow, interesting.  And another one just came in: 0a83443300000000.

That's pretty low volume, though, so I'll just look into the report to see what
I can learn and still remove the DumpWithoutCrashing.  As before, I'll leave a
DCHECK in its place.

-    NOTREACHED() << "Shouldn't set an empty PageState.";
-  }
-
   // The document_sequence_number and item_sequence_number recorded in the
   // FrameNavigationEntry should not differ from the one coming with the update,
   // since it must come from the same document. Do not update it if a difference