Helper script that will look try to figure out what string values are stale in our GRD files.

BUG=23550
TEST=none
Review URL: http://codereview.chromium.org/1631003

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@44078 0039d316-1c4b-4281-b951-d872f2087c98

1 files changed, 161 insertions, 0 deletions

diff --git a/chrome/tools/check_grd_for_unused_strings.py b/chrome/tools/check_grd_for_unused_strings.py
new file mode 100755
index 0000000..f66eccd
--- /dev/null
+++ b/chrome/tools/check_grd_for_unused_strings.py
@@ -0,0 +1,161 @@
+#!/usr/bin/python
+
+# Copyright (c) 2010 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.
+
+"""Without any args, this simply loads the IDs out of a bunch of the Chrome GRD
+files, and then checks the subset of the code that loads the strings to try
+and figure out what isn't in use any more.
+You can give paths to GRD files and source directories to control what is
+check instead.
+"""
+
+import os
+import re
+import sys
+import xml.sax
+
+# Extra messages along the way
+# 1 - Print ids that are found in sources but not in the found id set
+# 2 - Files that aren't processes (don't match the source name regex)
+DEBUG = 0
+
+
+class GrdIDExtractor(xml.sax.handler.ContentHandler):
+  """Extracts the IDs from messages in GRIT files"""
+  def __init__(self):
+    self.id_set_ = set()
+
+  def startElement(self, name, attrs):
+    if name == 'message':
+      self.id_set_.add(attrs['name'])
+
+  def allIDs(self):
+    """Return all the IDs found"""
+    return self.id_set_.copy()
+
+
+def CheckForUnusedGrdIDsInSources(grd_files, src_dirs):
+  """Will collect the message ids out of the given GRD files and then scan
+  the source directories to try and figure out what ids are not currently
+  being used by any source.
+
+  grd_files:
+    A list of GRD files to collect the ids from.
+  src_dirs:
+    A list of directories to walk looking for source files.
+  """
+  # Collect all the ids into a large map
+  all_ids = set()
+  file_id_map = {}
+  for y in grd_files:
+    handler = GrdIDExtractor()
+    xml.sax.parse(y, handler)
+    files_ids = handler.allIDs()
+    file_id_map[y] = files_ids
+    all_ids |= files_ids
+
+
+  # The regex that will be used to check sources
+  id_regex = re.compile('IDS_[A-Z0-9_]+')
+
+  # Make sure the regex matches every id found.
+  got_err = False
+  for x in all_ids:
+    match = id_regex.search(x)
+    if match is None:
+      print 'ERROR: "%s" did not match our regex' % (x)
+      got_err = True
+    if not match.group(0) is x:
+      print 'ERROR: "%s" did not fully match our regex' % (x)
+      got_err = True
+  if got_err:
+    return 1
+
+  # The regex for deciding what is a source file
+  src_regex = re.compile('\.(([chm])|(mm)|(cc)|(cp)|(cpp)|(xib))$')
+
+  ids_left = all_ids.copy()
+
+  # Scanning time.
+  for src_dir in src_dirs:
+    for root, dirs, files in os.walk(src_dir):
+      # Remove svn directories from recursion
+      if '.svn' in dirs:
+        dirs.remove('.svn')
+      for file in files:
+        if src_regex.search(file.lower()):
+          full_path = os.path.join(root, file)
+          src_file_contents = open(full_path).read()
+          for match in sorted(set(id_regex.findall(src_file_contents))):
+            if match in ids_left:
+              ids_left.remove(match)
+            if DEBUG:
+              if not match in all_ids:
+                print '%s had "%s", which was not in the found IDs' % \
+                  (full_path, match)
+        elif DEBUG > 1:
+          full_path = os.path.join(root, file)
+          print 'Skipping %s.' % (full_path)
+
+  # Anything left?
+  if len(ids_left) > 0:
+    print 'The following ids are in GRD files, but *appear* to be unused:'
+    for file_path, file_ids in file_id_map.iteritems():
+      missing = ids_left.intersection(file_ids)
+      if len(missing) > 0:
+        print '  %s:' % (file_path)
+        print '\n'.join('    %s' % (x) for x in sorted(missing))
+
+  return 0
+
+
+if __name__ == '__main__':
+  # script lives in src/chrome/tools
+  chrome_tools_dir = os.path.dirname(os.path.abspath(sys.argv[0]))
+  src_dir = os.path.dirname(os.path.dirname(chrome_tools_dir))
+
+  # Collect the args into the right buckets
+  src_dirs = []
+  grd_files = []
+  for arg in sys.argv[1:]:
+    if arg.lower().endswith('.grd'):
+      grd_files.append(arg)
+    else:
+      src_dirs.append(arg)
+
+  # If no GRD files were given, default them:
+  if len(grd_files) == 0:
+    app_res_dir = os.path.join(src_dir, 'app', 'resources')
+    chrome_dir = os.path.join(src_dir, 'chrome')
+    chrome_app_dir = os.path.join(chrome_dir, 'app')
+    chrome_app_res_dir = os.path.join(chrome_app_dir, 'resources')
+    grd_files = [
+      os.path.join(app_res_dir, 'app_locale_settings.grd'),
+      os.path.join(app_res_dir, 'app_resources.grd'),
+      os.path.join(app_res_dir, 'app_strings.grd'),
+      os.path.join(chrome_app_dir, 'chromium_strings.grd'),
+      os.path.join(chrome_app_dir, 'generated_resources.grd'),
+      os.path.join(chrome_app_dir, 'google_chrome_strings.grd'),
+      os.path.join(chrome_app_res_dir, 'locale_settings.grd'),
+      os.path.join(chrome_app_res_dir, 'locale_settings_linux.grd'),
+      os.path.join(chrome_app_res_dir, 'locale_settings_mac.grd'),
+      os.path.join(chrome_app_res_dir, 'locale_settings_win.grd'),
+      os.path.join(chrome_app_dir, 'theme', 'theme_resources.grd'),
+      os.path.join(chrome_dir, 'browser', 'browser_resources.grd'),
+      os.path.join(chrome_dir, 'common', 'common_resources.grd'),
+      os.path.join(chrome_dir, 'renderer', 'renderer_resources.grd'),
+    ]
+
+  # If no source directories were given, default them:
+  if len(src_dirs) == 0:
+    src_dirs = [
+      os.path.join(src_dir, 'app'),
+      os.path.join(src_dir, 'chrome'),
+      os.path.join(src_dir, 'views'),
+      # nsNSSCertHelper.cpp has a bunch of ids
+      os.path.join(src_dir, 'third_party', 'mozilla_security_manager'),
+    ]
+
+  sys.exit(CheckForUnusedGrdIDsInSources(grd_files, src_dirs))