Delete fullscreen/mouselock pref data.

chrome/browser/ui/webui/settings/site_settings_handler.cc

Index: chrome/browser/ui/webui/settings/site_settings_handler.cc
diff --git a/chrome/browser/ui/webui/settings/site_settings_handler.cc b/chrome/browser/ui/webui/settings/site_settings_handler.cc
index 160a7934fb6162d8caed0d5c8de8bb64e208ba7d..d4b3dd9171ace60011d5e39665c01e91be96a930 100644
--- a/chrome/browser/ui/webui/settings/site_settings_handler.cc
+++ b/chrome/browser/ui/webui/settings/site_settings_handler.cc
@@ -224,6 +224,12 @@ void SiteSettingsHandler::OnContentSettingChanged(
   if (!site_settings::HasRegisteredGroupName(content_type))
     return;
 
+  // These content types are deprecated and should not trigger a UI update.

[Matt Giuca, 2016/10/26 04:18:24]

I wish I didn't have to do this (it isn't strictly necessary), but the problem
is that without this, the test SiteSettingsHandlerTest.Incognito fails:
chrome/browser/ui/webui/settings/site_settings_handler_unittest.cc

Because it uses gmock to expect that the final call is onIncognitoStatusChanged,
and this breaks that assumption.

This is temporary; it will be deleted in a follow-up CL
(https://codereview.chromium.org/2419413002/) when I remove these enum values.

[stevenjb, 2016/10/26 19:31:00]

Could you add a TODO with a bug link? Generally modifying production code for
tests kind of sucks, so let's make sure this is well documented even if it's
going away almost immediately.

[Matt Giuca, 2016/10/26 23:26:47]

Done.

+  if (content_type == CONTENT_SETTINGS_TYPE_FULLSCREEN ||
+      content_type == CONTENT_SETTINGS_TYPE_MOUSELOCK) {
+    return;
+  }
+
   if (primary_pattern.ToString().empty()) {
     CallJavascriptFunction(
         "cr.webUIListenerCallback",