Replace HistoryQuickProvider protobuf-based caching with an SQLite-based database.

chrome/browser/history/in_memory_url_index.h

 // Copyright (c) 2012 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #ifndef CHROME_BROWSER_HISTORY_IN_MEMORY_URL_INDEX_H_
 #define CHROME_BROWSER_HISTORY_IN_MEMORY_URL_INDEX_H_
 
 #include <functional>
 #include <map>
 #include <set>
 #include <string>
 #include <vector>
 
 #include "base/basictypes.h"
 #include "base/file_path.h"
 #include "base/gtest_prod_util.h"
 #include "base/memory/ref_counted.h"
 #include "base/memory/weak_ptr.h"
+#include "base/observer_list.h"
 #include "base/string16.h"
+#include "base/threading/sequenced_worker_pool.h"
 #include "chrome/browser/autocomplete/autocomplete_match.h"
 #include "chrome/browser/autocomplete/history_provider_util.h"
 #include "chrome/browser/cancelable_request.h"
 #include "chrome/browser/history/history.h"
 #include "chrome/browser/history/history_types.h"
 #include "chrome/browser/history/in_memory_url_index_types.h"
 #include "chrome/browser/history/scored_history_match.h"
+#include "chrome/browser/history/url_index_private_data.h"
 #include "content/public/browser/notification_observer.h"
 #include "content/public/browser/notification_registrar.h"
 #include "sql/connection.h"
 
 class HistoryQuickProviderTest;
 class Profile;
 
 namespace base {
 class Time;
 }
 
-namespace in_memory_url_index {
-class InMemoryURLIndexCacheItem;
-}
-
 namespace history {
 
-namespace imui = in_memory_url_index;
-
-class HistoryDatabase;
-class URLIndexPrivateData;
+class InMemoryURLIndexObserver;
+class URLDatabase;
+struct URLsDeletedDetails;
+struct URLsModifiedDetails;
 struct URLVisitedDetails;
-struct URLsModifiedDetails;
-struct URLsDeletedDetails;
 
 // The URL history source.
 // Holds portions of the URL database in memory in an indexed form.  Used to
 // quickly look up matching URLs for a given query string.  Used by
 // the HistoryURLProvider for inline autocomplete and to provide URL
 // matches to the omnibox.
 //
 // Note about multi-byte codepoints and the data structures in the
 // InMemoryURLIndex class: One will quickly notice that no effort is made to
 // insure that multi-byte character boundaries are detected when indexing the
 // words and characters in the URL history database except when converting
 // URL strings to lowercase. Multi-byte-edness makes no difference when
 // indexing or when searching the index as the final filtering of results
 // is dependent on the comparison of a string of bytes, not individual
 // characters. While the lookup of those bytes during a search in the
 // |char_word_map_| could serve up words in which the individual char16
 // occurs as a portion of a composite character the next filtering step
 // will eliminate such words except in the case where a single character
 // is being searched on and which character occurs as the second char16 of a
 // multi-char16 instance.
-class InMemoryURLIndex : public content::NotificationObserver,
-                         public base::SupportsWeakPtr<InMemoryURLIndex> {
+class InMemoryURLIndex : public content::NotificationObserver {
  public:
-  // Defines an abstract class which is notified upon completion of restoring
-  // the index's private data either by reading from the cache file or by
-  // rebuilding from the history database.
-  class RestoreCacheObserver {
+  // Observer is used for blocking until the InMemoryURLIndex has finished
+  // loading. Usage typically follows this pattern:
+  //    InMemoryURLIndex::Observer observer(index); // Create observer.
+  //    MessageLoop::current()->Run();              // Blocks until loaded.
+  //
+  class Observer {
    public:
-    virtual ~RestoreCacheObserver();
+    explicit Observer(InMemoryURLIndex* index);
 
-    // Callback that lets the observer know that the restore operation has
-    // completed. |succeeded| indicates if the restore was successful. This is
-    // called on the UI thread.
-    virtual void OnCacheRestoreFinished(bool succeeded) = 0;
-  };
+    // Called when the InMemoryURLIndex has completed loading.
+    virtual void Loaded();
 
-  // Defines an abstract class which is notified upon completion of saving
-  // the index's private data to the cache file.
-  class SaveCacheObserver {
-   public:
-    virtual ~SaveCacheObserver();
+   private:
+    friend class InMemoryURLIndexBaseTest;
+    virtual ~Observer();
 
-    // Callback that lets the observer know that the save succeeded.
-    // This is called on the UI thread.
-    virtual void OnCacheSaveFinished(bool succeeded) = 0;
+    InMemoryURLIndex* index_;
+
+    DISALLOW_COPY_AND_ASSIGN(Observer);
   };
 
   // |profile|, which may be NULL during unit testing, is used to register for
   // history changes. |history_dir| is a path to the directory containing the
   // history database within the profile wherein the cache and transaction
   // journals will be stored. |languages| gives a list of language encodings by
   // which URLs and omnibox searches are broken down into words and characters.
   InMemoryURLIndex(Profile* profile,
                    const FilePath& history_dir,
                    const std::string& languages);
   virtual ~InMemoryURLIndex();
 
   // Opens and prepares the index of historical URL visits. If the index private
-  // data cannot be restored from its cache file then it is rebuilt from the
-  // history database.
-  void Init();
+  // data cannot be restored from its cache database then it is rebuilt from the
+  // history database. |disable_cache| causes the InMemoryURLIndex to not create
+  // or use its cache database.
+  void Init(bool disable_cache);
 
-  // Signals that any outstanding initialization should be canceled and
-  // flushes the cache to disk.
-  void ShutDown();
+  // Signals that any outstanding initialization should be canceled.
+  void Shutdown();
+
+  // Returns true if the index has been loaded or rebuilt and so is available
+  // for use.
+  bool index_available() const { return index_available_; }
 
   // Scans the history index and returns a vector with all scored, matching
   // history items. This entry point simply forwards the call on to the
   // URLIndexPrivateData class. For a complete description of this function
   // refer to that class.
   ScoredHistoryMatches HistoryItemsForTerms(const string16& term_string);
 
-  // Sets the optional observers for completion of restoral and saving of the
-  // index's private data.
-  void set_restore_cache_observer(
-      RestoreCacheObserver* restore_cache_observer) {
-    restore_cache_observer_ = restore_cache_observer;
-  }
-  void set_save_cache_observer(SaveCacheObserver* save_cache_observer) {
-    save_cache_observer_ = save_cache_observer;
-  }
+  // Deletes the index entry, if any, for the given |url|.
+  void DeleteURL(const GURL& url);
 
  private:
   friend class ::HistoryQuickProviderTest;
+  friend class InMemoryURLIndex::Observer;
+  friend class InMemoryURLIndexCacheTest;
   friend class InMemoryURLIndexTest;
-  friend class InMemoryURLIndexCacheTest;
+  friend class InMemoryURLIndexBaseTest;
+  friend class IntercessionaryIndexTest;
+  FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, ExpireRow);
+  FRIEND_TEST_ALL_PREFIXES(IntercessionaryIndexTest, CacheDatabaseFailure);
+  FRIEND_TEST_ALL_PREFIXES(IntercessionaryIndexTest,
+                           ShutdownDuringCacheRefresh);
   FRIEND_TEST_ALL_PREFIXES(LimitedInMemoryURLIndexTest, Initialization);
 
-  // Creating one of me without a history path is not allowed (tests excepted).
-  InMemoryURLIndex();
-
   // HistoryDBTask used to rebuild our private data from the history database.
   class RebuildPrivateDataFromHistoryDBTask : public HistoryDBTask {
    public:
-    explicit RebuildPrivateDataFromHistoryDBTask(
-        InMemoryURLIndex* index,
-        const std::string& languages,
-        const std::set<std::string>& scheme_whitelist);
+    explicit RebuildPrivateDataFromHistoryDBTask(InMemoryURLIndex* index);
 
     virtual bool RunOnDBThread(HistoryBackend* backend,
                                history::HistoryDatabase* db) OVERRIDE;
     virtual void DoneRunOnMainThread() OVERRIDE;
 
    private:
     virtual ~RebuildPrivateDataFromHistoryDBTask();
 
     InMemoryURLIndex* index_;  // Call back to this index at completion.
-    std::string languages_;  // Languages for word-breaking.
-    std::set<std::string> scheme_whitelist_;  // Schemes to be indexed.
     bool succeeded_;  // Indicates if the rebuild was successful.
     scoped_refptr<URLIndexPrivateData> data_;  // The rebuilt private data.
 
     DISALLOW_COPY_AND_ASSIGN(RebuildPrivateDataFromHistoryDBTask);
   };
 
-  // Initializes all index data members in preparation for restoring the index
-  // from the cache or a complete rebuild from the history database.
-  void ClearPrivateData();
+  // For unit testing only.
+  InMemoryURLIndex(const FilePath& history_dir, const std::string& languages);
 
-  // Constructs a file path for the cache file within the same directory where
-  // the history database is kept and saves that path to |file_path|. Returns
-  // true if |file_path| can be successfully constructed. (This function
-  // provided as a hook for unit testing.)
-  bool GetCacheFilePath(FilePath* file_path);
+  // Completes index initialization once the cache DB has been initialized.
+  void OnPrivateDataInitDone(bool succeeded);
 
-  // Restores the index's private data from the cache file stored in the
-  // profile directory.
-  void PostRestoreFromCacheFileTask();
+  // Adds or removes an observer that is notified when the index has been
+  // loaded.
+  void AddObserver(InMemoryURLIndex::Observer* observer);
+  void RemoveObserver(InMemoryURLIndex::Observer* observer);
+
+  // Handles notifications of history changes.
+  virtual void Observe(int notification_type,
+                       const content::NotificationSource& source,
+                       const content::NotificationDetails& details) OVERRIDE;
+
+  // Notification handlers.
+  void OnURLVisited(const URLVisitedDetails* details);
+  void OnURLsModified(const URLsModifiedDetails* details);
+  void OnURLsDeleted(const URLsDeletedDetails* details);
+
+  // Posts any outstanding updates to the index which were queued while the
+  // index was being initialized.
+  void FlushPendingUpdates();
+
+  // Restores the index's private data from the cache database stored in the
+  // profile directory. If no database is found or can be restored then look
+  // for an old version protobuf-based cache file.
+  void PostRestoreFromCacheTask();
+
+  // Determines if the private data was successfully restored from the cache
+  // database, as indicated by |succeeded|, or if the private data must be
+  // rebuilt from the history database. If successful, notifies any
+  // |restore_cache_observer_|. Otherwise, kicks off a rebuild from the history
+  // database.
+  void OnCacheRestoreDone(bool succeeded);
+
+  // Notifies all observers that the index has been loaded or rebuilt.
+  void NotifyHasLoaded();
+
+  // Rebuilds the index from the history database if the history database has
+  // been loaded, otherwise registers for the history loaded notification so
+  // that the rebuild can take place at a later time.
+  void RebuildFromHistoryIfLoaded();
 
   // Schedules a history task to rebuild our private data from the history
   // database.
   void ScheduleRebuildFromHistory();
 
   // Callback used by RebuildPrivateDataFromHistoryDBTask to signal completion
   // or rebuilding our private data from the history database. |succeeded|
   // will be true if the rebuild was successful. |data| will point to a new
   // instanceof the private data just rebuilt.
   void DoneRebuidingPrivateDataFromHistoryDB(
       bool succeeded,
       scoped_refptr<URLIndexPrivateData> private_data);
 
-  // Rebuilds the history index from the history database in |history_db|.
-  // Used for unit testing only.
-  void RebuildFromHistory(HistoryDatabase* history_db);
+  // Posts a task to completely reset the private data and the backing cache.
+  void PostResetPrivateDataTask();
 
-  // Determines if the private data was successfully reloaded from the cache
-  // file or if the private data must be rebuilt from the history database.
-  // |private_data_ptr|'s data will be NULL if the cache file load failed. If
-  // successful, sets the private data and notifies any
-  // |restore_cache_observer_|. Otherwise, kicks off a rebuild from the history
-  // database.
-  void OnCacheLoadDone(
-    scoped_refptr<URLIndexPrivateData> private_data_ptr);
+  // Posts a task to completely replace the cache database with a current
+  // image of the index private data.
+  void PostRefreshCacheTask();
 
-  // Callback function that sets the private data from the just-restored-from-
-  // file |private_data|. Notifies any |restore_cache_observer_| that the
-  // restore has succeeded.
-  void OnCacheRestored(URLIndexPrivateData* private_data);
+  // Callback used by PostRefreshCacheTask and PostResetPrivateDataTask to
+  // notify observers that the cache database contents have been refreshed or
+  // reset and that the loading of the index is complete.
+  void OnCacheRefreshOrResetDone();
 
-  // Posts a task to cache the index private data and write the cache file to
-  // the profile directory.
-  void PostSaveToCacheFileTask();
+  // Attempts to refresh the cache database in response to a notification that
+  // an update transaction has failed. If the refresh fails then the cache
+  // database is ignored and an attempt will be made to rebuild the cache
+  // the next time the associated profile is opened.
+  void RepairCacheDatabase();
 
-  // Saves private_data_ to the given |path|. Runs on the UI thread.
-  // Provided for unit testing so that a test cache file can be used.
-  void DoSaveToCacheFile(const FilePath& path);
+  // Returns a pointer to our private data.
+  scoped_refptr<URLIndexPrivateData> private_data() { return private_data_; }
 
-  // Notifies the observer, if any, of the success of the private data caching.
-  // |succeeded| is true on a successful save.
-  void OnCacheSaveDone(scoped_refptr<RefCountedBool> succeeded);
-
-  // Handles notifications of history changes.
-  virtual void Observe(int notification_type,
-                       const content::NotificationSource& source,
-                       const content::NotificationDetails& details) OVERRIDE;
-
-  // Notification handlers.
-  void OnURLVisited(const URLVisitedDetails* details);
-  void OnURLsModified(const URLsModifiedDetails* details);
-  void OnURLsDeleted(const URLsDeletedDetails* details);
-
-  // Sets the directory wherein the cache file will be maintained.
-  // For unit test usage only.
-  void set_history_dir(const FilePath& dir_path) { history_dir_ = dir_path; }
-
-  // Returns a pointer to our private data. For unit testing only.
-  URLIndexPrivateData* private_data() { return private_data_.get(); }
-
-  // Returns the set of whitelisted schemes. For unit testing only.
-  const std::set<std::string>& scheme_whitelist() { return scheme_whitelist_; }
+  // Returns the blocking pool sequence token.
+  base::SequencedWorkerPool::SequenceToken sequence_token_for_testing() {
+    return sequence_token_;
+  }
 
   // The profile, may be null when testing.
   Profile* profile_;
 
-  // Directory where cache file resides. This is, except when unit testing,
-  // the same directory in which the profile's history database is found. It
-  // should never be empty.
-  FilePath history_dir_;
-
   // Languages used during the word-breaking process during indexing.
   std::string languages_;
 
-  // Only URLs with a whitelisted scheme are indexed.
-  std::set<std::string> scheme_whitelist_;
+  // Directory where cache database or protobuf-based cache file resides.
+  // This is, except when unit testing, the same directory in which the
+  // profile's history database is found.
+  FilePath history_dir_;
 
   // The index's durable private data.
   scoped_refptr<URLIndexPrivateData> private_data_;
 
-  // Observers to notify upon restoral or save of the private data cache.
-  RestoreCacheObserver* restore_cache_observer_;
-  SaveCacheObserver* save_cache_observer_;
+  // Sequence token for coordinating database tasks. This is shared with
+  // our private data and its cache database.
+  const base::SequencedWorkerPool::SequenceToken sequence_token_;
+
+  bool index_available_;  // True when index is available for updating.
+
+  // Contains index updates queued up while the index is unavailable. This
+  // usually during profile startup.
+  enum UpdateType { UPDATE_VISIT, DELETE_VISIT };
+
+  struct IndexUpdateItem {
+    IndexUpdateItem(UpdateType update_type, URLRow row);
+    ~IndexUpdateItem();
+
+    UpdateType update_type;
+    URLRow row;   // The row to be updated or deleted.
+  };
+  typedef std::vector<IndexUpdateItem> PendingUpdates;
+  PendingUpdates pending_updates_;
+
+  ObserverList<InMemoryURLIndex::Observer> observers_;
 
   CancelableRequestConsumer cache_reader_consumer_;
   content::NotificationRegistrar registrar_;
 
   // Set to true once the shutdown process has begun.
   bool shutdown_;
 
-  // Set to true when changes to the index have been made and the index needs
-  // to be cached. Set to false when the index has been cached. Used as a
-  // temporary safety check to insure that the cache is saved before the
-  // index has been destructed.
-  // TODO(mrossetti): Eliminate once the transition to SQLite has been done.
-  // http://crbug.com/83659
-  bool needs_to_be_cached_;
+  // Note: This should remain the last member so it'll be destroyed and
+  // invalidate its weak pointers before any other members are destroyed.
+  base::WeakPtrFactory<InMemoryURLIndex> weak_ptr_factory_;
 
   DISALLOW_COPY_AND_ASSIGN(InMemoryURLIndex);
 };
 
 }  // namespace history
 
 #endif  // CHROME_BROWSER_HISTORY_IN_MEMORY_URL_INDEX_H_