[StorageMonitor] Move StorageMonitor ownership to BrowserProcessImpl

chrome/browser/storage_monitor/storage_monitor.h

Index: chrome/browser/storage_monitor/storage_monitor.h
diff --git a/chrome/browser/storage_monitor/storage_monitor.h b/chrome/browser/storage_monitor/storage_monitor.h
index 5661c80fb5e9252682528a16d2c4038105f2154a..7d2dd8e0ddc57d4792bac423ac8b8675700b8a2d 100644
--- a/chrome/browser/storage_monitor/storage_monitor.h
+++ b/chrome/browser/storage_monitor/storage_monitor.h
@@ -68,10 +68,17 @@ class StorageMonitor {
     EJECT_FAILURE
   };
 
-  // Returns a pointer to an object owned by the BrowserMainParts, with lifetime
-  // somewhat shorter than a process Singleton.
+  // Returns a pointer to a created per-platform object with the StorageMonitor
+  // interface.

[vandebo (ex-Chrome), 2013/07/10 16:57:45]

Should this only be called by BrowserProcess (especially with respect to the
following comment?)  If that's the case, then maybe Create should be private
with BrowserProcess as a friend?

[Greg Billock, 2013/07/10 21:57:17]

Hmmm. Could do.

+  static StorageMonitor* Create();
+
+  // Returns a pointer to an object owned by BrowserProcess, with lifetime
+  // starting before main message loop start, and ending after main message loop
+  // shutdown.

[vandebo (ex-Chrome), 2013/07/10 16:57:45]

Outside that lifetime this returns Null?

[Greg Billock, 2013/07/10 21:11:16]

Added comment

   static StorageMonitor* GetInstance();
 
+  virtual ~StorageMonitor();
+
   // Ensures that the storage monitor is initialized. The provided callback, If
   // non-null, will be called when initialization is complete. If initialization
   // has already completed, this callback will be invoked within the calling
@@ -132,11 +139,6 @@ class StorageMonitor {
   friend class ::SystemInfoStorageApiTest;
 
   StorageMonitor();
-  virtual ~StorageMonitor();
-
-  // Removes the existing singleton for testing.
-  // (So that a new one can be created.)
-  static void RemoveSingletonForTesting();
 
   virtual Receiver* receiver() const;

[vandebo (ex-Chrome), 2013/07/10 16:57:45]

aside: weren't we going to get rid of this method once we passed receiver_ into
the constructor of all of the implementations?

[Greg Billock, 2013/07/10 21:11:16]

I think so. There are still various tests that use it. Still some work to do
there...