Add deduplication logic to .pak files

build/android/resource_sizes.py

 #!/usr/bin/env python
 # Copyright (c) 2011 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.
 
 """Prints the size of each given file and optionally computes the size of
    libchrome.so without the dependencies added for building with android NDK.
    Also breaks down the contents of the APK to determine the installed size
    and assign size contributions to different classes of file.
 """
@@ skipping 544 matching lines @@
         100.0 * pak.compress_size / total_compress_size,
         _FormatBytes(pak.file_size),
         100.0 * pak.file_size / total_file_size)
 
   print
   print 'Analyzing pak resources in %s...' % apk_filename
 
   # Calculate aggregate stats about resources across pak files.
   resource_count_map = collections.defaultdict(int)
   resource_size_map = collections.defaultdict(int)
+  seen_data_ids = set()
+  alias_overhead_bytes = 4
   resource_overhead_bytes = 6
   for pak in paks:
-    for r in pak.resources:
-      resource_count_map[r] += 1
-      resource_size_map[r] += len(pak.resources[r]) + resource_overhead_bytes
-
+    for k, v in pak.resources.iteritems():
+      resource_count_map[k] += 1
+      if id(v) not in seen_data_ids:
+        seen_data_ids.add(id(v))
+        resource_size_map[k] += resource_overhead_bytes + len(v)
+      else:
+        resource_size_map[k] += alias_overhead_bytes
   # Output the overall resource summary.
   total_resource_size = sum(resource_size_map.values())
   total_resource_count = len(resource_count_map)
   assert total_resource_size <= total_file_size
   print 'Total pak resources: %s' % total_resource_count
   print 'Total uncompressed resource size: %s' % _FormatBytes(
       total_resource_size)
   print
 
   resource_id_name_map, resources_id_header_map = _AnnotatePakResources()
@@ skipping 251 matching lines @@
         chartjson=chartjson)
   if chartjson:
     results_path = os.path.join(args.output_dir, 'results-chart.json')
     logging.critical('Dumping json to %s', results_path)
     with open(results_path, 'w') as json_file:
       json.dump(chartjson, json_file)
 
 
 if __name__ == '__main__':
   sys.exit(main())