Issue 2325 - Add a way to set settings in libadblockplus for FRP and automatic updates

test/Prefs.cpp

Index: test/Prefs.cpp
===================================================================
--- a/test/Prefs.cpp
+++ b/test/Prefs.cpp
@@ -78,14 +78,16 @@
       Reset();
     }
 
-    void Reset()
+    void Reset(std::map<std::string, AdblockPlus::JsValuePtr> preconfiguredPrefs =
+               std::map<std::string, AdblockPlus::JsValuePtr>())
     {
       jsEngine = AdblockPlus::JsEngine::New();
       jsEngine->SetLogSystem(AdblockPlus::LogSystemPtr(new LazyLogSystem));
       jsEngine->SetFileSystem(fileSystemPtr);
       jsEngine->SetWebRequest(AdblockPlus::WebRequestPtr(new LazyWebRequest));
 
-      filterEngine.reset(new AdblockPlus::FilterEngine(jsEngine));
+      filterEngine.reset(
+        new AdblockPlus::FilterEngine(jsEngine, preconfiguredPrefs));
     }
   };
 }
@@ -147,3 +149,16 @@
 
   ASSERT_EQ(24, filterEngine->GetPref("patternsbackupinterval")->AsInt());
 }
+
+TEST_F(PrefsTest, PreconfiguredPrefs)
+{
+  std::map<std::string, AdblockPlus::JsValuePtr> preconfiguredPrefs;
+  preconfiguredPrefs["disable_auto_updates"] = jsEngine->NewValue(false);
+  preconfiguredPrefs["suppress_first_run_page"] = jsEngine->NewValue(true);
+  preconfiguredPrefs["unsupported_preconfig"] = jsEngine->NewValue(true);
+  Reset(preconfiguredPrefs);
+
+  ASSERT_FALSE(filterEngine->GetPref("disable_auto_updates")->AsBool());
+  ASSERT_TRUE(filterEngine->GetPref("suppress_first_run_page")->AsBool());
+  ASSERT_FALSE(filterEngine->GetPref("unsupported_preconfig")->AsBool());

[Eric, 2015/06/10 17:48:16]

This assertion should really be in a separate test, since it's exercising a
different piece of the code.

[Oleksandr, 2015/06/12 10:47:57]

Done.

+}

[Eric, 2015/06/10 17:48:16]

With preconfigurable preferences, we now have three different sources of
preferences: (a) defaults, (b) preconfigured values, (c) user-set values. We
need tests for the different combinations of these.

The existing test exercises (a) vs. (b). Since (a) is always present as a
source, there are three more combinations that need testing.
-- (a) only
-- (a) and (c)
-- all of (a), (b), and (c)

[Oleksandr, 2015/06/12 07:36:14]

I think (a) and (a) and (c) are tested quite well in other tests in this file

[Felix Dahlke, 2015/06/12 07:55:03]

Most of that is covered here, but I think we should have a test for the b/c
combination, in particular when it comes to pref loading. Your earlier patch
wouldn't have passed that test, so I think it's worth having :)

Basically something like PrefsPersistWhenPreconfigured or something, similar to
the PrefsPersist test above:

1. Preconfigure a pref
2. Ensure it's set properly
3. Set it to a different value
4. Ensure it's set properly
5. Reload prefs (how PrefsPersist does it)
6. Ensure it's still set to the user defined value

[Oleksandr, 2015/06/12 10:47:57]

Done. That's a great test, actually. Revealed one more bug, which is fixed in
the latest patchset.