Avoid fragmenting the heap too much while reconstructing the index.

net/disk_cache/simple/simple_index_file.h

Index: net/disk_cache/simple/simple_index_file.h
diff --git a/net/disk_cache/simple/simple_index_file.h b/net/disk_cache/simple/simple_index_file.h
index b536df9a1e72cbaaa8a32f01ef33e82fa6f915d3..848988672644b9760e30891fa9030e2ac82b886a 100644
--- a/net/disk_cache/simple/simple_index_file.h
+++ b/net/disk_cache/simple/simple_index_file.h
@@ -15,6 +15,7 @@
 #include "base/logging.h"
 #include "base/memory/scoped_ptr.h"
 #include "base/pickle.h"
+#include "base/platform_file.h"

[gavinp, 2013/08/21 11:49:47]

Pourquoi? Just forward declare base::PlatformFileInfo I say!

[Philippe, 2013/08/21 12:41:38]

Yeah my bad. I used to have GetPlatformFileInfo() return a value-typed
PlatformFileInfo.

 #include "base/port.h"
 #include "net/base/net_export.h"
 #include "net/disk_cache/simple/simple_index.h"
@@ -97,6 +98,9 @@ class NET_EXPORT_PRIVATE SimpleIndexFile {
  private:
   friend class WrappedSimpleIndexFile;
 
+  // Used for cache directory traversal.
+  typedef base::Callback<void (const char* file_name)> EntryFileCallback;
+
   // When loading the entries from disk, add this many extra hash buckets to
   // prevent reallocation on the IO thread when merging in new live entries.
   static const int kExtraSizeForMerge = 512;
@@ -120,9 +124,32 @@ class NET_EXPORT_PRIVATE SimpleIndexFile {
 
   // Given the contents of an index file |data| of length |data_len|, returns
   // the corresponding EntrySet. Returns NULL on error.
-  static void Deserialize(const char* data, int data_len,
+  static void Deserialize(const char* data,
+                          int data_len,
                           SimpleIndexLoadResult* out_result);
 
+  // Implemented either in simple_index_file_posix.cc or
+  // simple_index_file_win.cc. base::FileEnumerator turned out to be very
+  // expensive in terms of memory usage therefore it's used only on non-POSIX
+  // environments for convenience (for now). Note that this function is
+  // synchronous although its signature could suggest otherwise.  Returns
+  // whether the traversal succeeded.
+  static bool TraverseCacheDirectory(
+      const std::string& cache_path,
+      const EntryFileCallback& entry_file_callback);
+
+  // Returns whether file info was successfully retrieved.
+  static bool GetPlatformFileInfo(const char* path,
+                                  base::PlatformFileInfo* info);
+
+  // Called for each cache directory traversal iteration. The provided
+  // non-cleared |buffer| can be used accross iterations to avoid doing too many
+  // heap allocations given that cache directories can be very large.
+  static void ProcessEntryFile(const std::string& cache_path,

[gavinp, 2013/08/21 11:49:47]

Can this method be moved into an anonymous namespace in simple_index_file.cc
easily?

[Philippe, 2013/08/21 12:41:38]

Yeah good point.

+                               std::vector<char>* buffer,

[gavinp, 2013/08/21 11:49:47]

char*! Let your inner C programmer out.

[Philippe, 2013/08/21 12:41:38]

Ahaha :) I got rid of the buffer.

+                               SimpleIndex::EntrySet* entries,
+                               const char* file_name);
+
   // Scan the index directory for entries, returning an EntrySet of all entries
   // found.
   static void SyncRestoreFromDisk(const base::FilePath& cache_directory,