Updates WKWebViewSecurityUtil to handle deprecated enum values.

ios/web/web_state/wk_web_view_security_util.mm

Index: ios/web/web_state/wk_web_view_security_util.mm
diff --git a/ios/web/web_state/wk_web_view_security_util.mm b/ios/web/web_state/wk_web_view_security_util.mm
index b927e676edfcd85b8896d800a6c88d512afb15ad..d186debe1faec2ba3d1620cfe59194241d0ca5d8 100644
--- a/ios/web/web_state/wk_web_view_security_util.mm
+++ b/ios/web/web_state/wk_web_view_security_util.mm
@@ -139,9 +139,15 @@ SecurityStyle GetSecurityStyleFromTrustResult(SecTrustResultType result) {
     case kSecTrustResultFatalTrustFailure:
     case kSecTrustResultOtherError:
       return SECURITY_STYLE_AUTHENTICATION_BROKEN;
+
+    // TODO(crbug.com/): This default clause exists because
+    // kSecTrustResultConfirm was deprecated in iOS7, but leads to a compile
+    // error if used with newer SDKs.  Remove the default clause once this
+    // switch statement successfully compiles without kSecTrustResultConfirm.
+    default:

[rohitrao (ping after 24h), 2016/06/14 15:20:32]

Eugene asked whether this should be hidden behind a preprocessor guard.  My
feeling is that the iOS9 SDK is never going to change to include new enum
values, and anything newer will have a default clause, so adding a guard doesn't
meaningfully protect us from anything.  Thoughts?

If we're ok with this approach, I'll file a bug for the TODO.

[Eugene But (OOO till 7-30), 2016/06/14 15:35:07]

I'm totally fine with having default. And I'm not even sure if we ned a TODO
here. Having a comment is fine IMO.

[sdefresne, 2016/06/14 18:03:09]

+1 to eugenebut comment.

+      NOTREACHED();
+      return SECURITY_STYLE_UNKNOWN;
   }
-  NOTREACHED();
-  return SECURITY_STYLE_UNKNOWN;
 }
 
 }  // namespace web