Use GetPermissionStatus instead of DesktopNotificationProfileUtil::GetContentSetting

chrome/browser/android/preferences/website_preference_bridge.cc

Index: chrome/browser/android/preferences/website_preference_bridge.cc
diff --git a/chrome/browser/android/preferences/website_preference_bridge.cc b/chrome/browser/android/preferences/website_preference_bridge.cc
index 62f8784ad826e58770d414d87097f71c5e294a26..d435c71a2f1009e1f0618042a2143c2c3a5ad440 100644
--- a/chrome/browser/android/preferences/website_preference_bridge.cc
+++ b/chrome/browser/android/preferences/website_preference_bridge.cc
@@ -315,9 +315,8 @@ static jint GetNotificationSettingForOrigin(
     const JavaParamRef<jstring>& origin,
     const JavaParamRef<jstring>& embedder,
     jboolean is_incognito) {
-  return DesktopNotificationProfileUtil::GetContentSetting(
-      GetActiveUserProfile(is_incognito),
-      GURL(ConvertJavaStringToUTF8(env, origin)));
+  return GetSettingForOrigin(env, CONTENT_SETTINGS_TYPE_NOTIFICATIONS,
+                             origin, origin, is_incognito);

[Bernhard Bauer, 2016/06/02 11:19:01]

Is it on purpose that we ignore the embedder and use the origin instead?

[Michael van Ouwerkerk, 2016/06/02 12:03:01]

Yes, that's on purpose: https://codereview.chromium.org/921803002/

[lshang, 2016/06/02 12:05:25]

I use origins for both params here mainly because the previous code
DesktopNotificationProfileUtil::GetContentSetting() only uses origins, and I'm
making them consistent before and after the migration to introduce bugs. 

On more thoughts, CONTENT_SETTINGS_TYPE_NOTIFICATIONS is
REQUESTING_ORIGIN_ONLY_SCOPE, and it only uses origin when setting content
settings. So I think notifications only cares about origin.

[Bernhard Bauer, 2016/06/02 12:23:04]

Why don't we get rid of the parameter then?

 }
 
 static void SetNotificationSettingForOrigin(