Reland of Enable whitelist generation for official builds.

tools/resources/generate_resource_whitelist.py

Index: tools/resources/generate_resource_whitelist.py
diff --git a/tools/resources/find_used_resources.py b/tools/resources/generate_resource_whitelist.py
similarity index 59%
rename from tools/resources/find_used_resources.py
rename to tools/resources/generate_resource_whitelist.py
index 0528115ead59cfadbc1d16cb2b38090e48937a7a..527ceb484158c9fcf32cb5d9dc34e7561d7471d9 100755
--- a/tools/resources/find_used_resources.py
+++ b/tools/resources/generate_resource_whitelist.py
@@ -1,5 +1,5 @@
 #!/usr/bin/env python
-# Copyright 2014 The Chromium Authors. All rights reserved.
+# Copyright 2016 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.
 
@@ -8,18 +8,22 @@ import os
 import re
 import sys
 
-USAGE = """find_used_resources.py [-h] [-i INPUT] [-o OUTPUT]
+USAGE = """generate_resource_whitelist.py [-h] [-i INPUT] [-o OUTPUT]
 
-Outputs the sorted list of resource ids that are part of unknown pragma warning
-in the given build log.
+INPUT specifies a file containing existing resource IDs that should be
+whitelisted, where each line of INPUT contains a single resource ID.
+
+Creates a resource whitelist by collecting existing resource IDs that are part
+of unknown pragma warnings and adds additional arch specfic resource IDs.
 
 This script is used to find the resources that are actually compiled in Chrome
-in order to only include the needed strings/images in Chrome PAK files. The
-script parses out the list of used resource ids. These resource ids show up in
-the build output after building Chrome with gyp variable
-enable_resource_whitelist_generation set to 1. This gyp flag causes the compiler
-to print out a UnknownPragma message every time a resource id is used. E.g.:
-foo.cc:22:0: warning: ignoring #pragma whitelisted_resource_12345
+in order to only include the needed strings/images in Chrome PAK files.
+These resource IDs show up in the build output after building Chrome with
+gn variable enable_resource_whitelist_generation set to true.
+This causes the compiler to print out an UnknownPragma message every time a
+resource ID is used.
+
+E.g. foo.cc:22:0: warning: ignoring #pragma whitelisted_resource_12345
 [-Wunknown-pragmas]
 
 On Windows, the message is simply a message via __pragma(message(...)).
@@ -35,7 +39,8 @@ ARCH_SPECIFIC_RESOURCES = [
   'IDS_VERSION_UI_32BIT',
 ]
 
-def FindResourceIds(header, resource_names):
+
+def _FindResourceIds(header, resource_names):
   """Returns the numerical resource IDs that correspond to the given resource
      names, as #defined in the given header file."
   """
@@ -49,36 +54,24 @@ def FindResourceIds(header, resource_names):
                     ', '.join(str(i) for i in res_ids))
   return set(res_ids)
 
-def GetResourceIdsInPragmaWarnings(input):
-  """Returns set of resource ids that are inside unknown pragma warnings
-     for the given input.
-  """
-  used_resources = set()
-  unknown_pragma_warning_pattern = re.compile(
-      'whitelisted_resource_(?P<resource_id>[0-9]+)')
-  for ln in input:
-    match = unknown_pragma_warning_pattern.search(ln)
-    if match:
-      resource_id = int(match.group('resource_id'))
-      used_resources.add(resource_id)
-  return used_resources
-
-def Main():
+
+def main():
   parser = argparse.ArgumentParser(usage=USAGE)
   parser.add_argument(
       '-i', '--input', type=argparse.FileType('r'), default=sys.stdin,
-      help='The build log to read (default stdin)')
+      help='A resource whitelist where each line contains one resource ID')
   parser.add_argument(
       '-o', '--output', type=argparse.FileType('w'), default=sys.stdout,
       help='The resource list path to write (default stdout)')
-  parser.add_argument('--out-dir', required=True,
+  parser.add_argument(
+      '--out-dir', required=True,
       help='The out target directory, for example out/Release')
 
   args = parser.parse_args()
 
-
-  used_resources = GetResourceIdsInPragmaWarnings(args.input)
-  used_resources |= FindResourceIds(
+  used_resources = set()
+  used_resources.update([int(resource_id) for resource_id in args.input])
+  used_resources |= _FindResourceIds(
       os.path.join(args.out_dir, COMPONENTS_STRINGS_HEADER),
       ARCH_SPECIFIC_RESOURCES)
 
@@ -86,4 +79,4 @@ def Main():
     args.output.write('%d\n' % resource_id)
 
 if __name__ == '__main__':
-  Main()
+  main()