Add whitelist support to grit.

- Support -w path to add whitelists to grit (and grit_info).
- Use the whitelist to tag what items should be skipped when generating output.

A whitelist is used instead of a blacklist, since there has been some other
work into generating what resources are really needed by the build.

Another option instead of whitelists would have been to add yet more conditions
to the grd files, but they are already getting pretty complex with the
conditions already in the files.  And looking at trying to do conditions based
on features not only made the files more complex, but also doesn't handle the
problem when a single feature has some strings that are only valid for a subset
of the supported platforms.
Review URL: http://codereview.chromium.org/6821056

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@81393 0039d316-1c4b-4281-b951-d872f2087c98

4 files changed, 82 insertions, 10 deletions

diff --git a/tools/grit/grit/format/data_pack.py b/tools/grit/grit/format/data_pack.py
index 20a8f73..c8a46ad 100755
--- a/tools/grit/grit/format/data_pack.py
+++ b/tools/grit/grit/format/data_pack.py
@@ -1,5 +1,5 @@
 #!/usr/bin/python2.4
-# Copyright (c) 2009 The Chromium Authors. All rights reserved.
+# 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.
 
@@ -41,7 +41,10 @@ class DataPack(interface.ItemFormatter):
       return nodes
     if (isinstance(item, include.IncludeNode) or
         isinstance(item, message.MessageNode)):
-      return [item]
+      # Include this node if it wasn't marked as skipped by a whitelist.
+      if not item.WhitelistMarkedAsSkip():
+        return [item]
+      return nodes
     for child in item.children:
       nodes.extend(DataPack.GetDataNodes(child))
     return nodes
diff --git a/tools/grit/grit/node/base.py b/tools/grit/grit/node/base.py
index b35c200..631733a 100644
--- a/tools/grit/grit/node/base.py
+++ b/tools/grit/grit/node/base.py
@@ -1,5 +1,5 @@
 #!/usr/bin/python2.4
-# Copyright (c) 2006-2008 The Chromium Authors. All rights reserved.
+# 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.
 
@@ -26,6 +26,9 @@ class Node(grit.format.interface.ItemFormatter):
   _CONTENT_TYPE_CDATA = 1  # Only CDATA, no children.
   _CONTENT_TYPE_MIXED = 2  # CDATA and children, possibly intermingled
 
+  # Default nodes to not whitelist skipped
+  _whitelist_marked_as_skip = False
+
   def __init__(self):
     self.children = []        # A list of child elements
     self.mixed_content = []   # A list of u'' and/or child elements (this
@@ -520,6 +523,18 @@ class Node(grit.format.interface.ItemFormatter):
       p = p.parent
     return False
 
+  def WhitelistMarkedAsSkip(self):
+    '''Returns true if the node is marked to be skipped in the output by a
+    whitelist.
+    '''
+    return self._whitelist_marked_as_skip
+
+  def SetWhitelistMarkedAsSkip(self, mark_skipped):
+    '''Sets WhitelistMarkedAsSkip.
+    '''
+    self._whitelist_marked_as_skip = mark_skipped
+
+
 class ContentNode(Node):
   '''Convenience baseclass for nodes that can have content.'''
   def _ContentType(self):
diff --git a/tools/grit/grit/tool/build.py b/tools/grit/grit/tool/build.py
index 4b7ddd9..635cc41 100644
--- a/tools/grit/grit/tool/build.py
+++ b/tools/grit/grit/tool/build.py
@@ -1,5 +1,5 @@
 #!/usr/bin/python2.4
-# Copyright (c) 2006-2008 The Chromium Authors. All rights reserved.
+# 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.
 
@@ -61,6 +61,9 @@ Options:
                     resources_ids next to grit.py.  Set to an empty string
                     if you don't want to use a first id file.
 
+  -w WHITELISTFILE  Path to a file containing the string names of the
+                    resources to include.  Anything not listed is dropped.
+
 
 Conditional inclusion of resources only affects the output of files which
 control which resources get linked into a binary, e.g. it affects .rc files
@@ -75,7 +78,8 @@ are exported to translation interchange files (e.g. XMB files), etc.
   def Run(self, opts, args):
     self.output_directory = '.'
     first_id_filename = None
-    (own_opts, args) = getopt.getopt(args, 'o:D:E:f:')
+    whitelist_filenames = []
+    (own_opts, args) = getopt.getopt(args, 'o:D:E:f:w:')
     for (key, val) in own_opts:
       if key == '-o':
         self.output_directory = val
@@ -87,6 +91,8 @@ are exported to translation interchange files (e.g. XMB files), etc.
         os.environ[env_name] = env_value
       elif key == '-f':
         first_id_filename = val
+      elif key == '-w':
+        whitelist_filenames.append(val)
 
     if len(args):
       print "This tool takes no tool-specific arguments."
@@ -98,6 +104,15 @@ are exported to translation interchange files (e.g. XMB files), etc.
       self.VerboseOut('Output directory: %s (absolute path: %s)\n' %
                       (self.output_directory,
                        os.path.abspath(self.output_directory)))
+
+    if whitelist_filenames:
+      self.whitelist_names = set()
+      for whitelist_filename in whitelist_filenames:
+        self.VerboseOut('Using whitelist: %s\n' % whitelist_filename);
+        whitelist_file = open(whitelist_filename)
+        self.whitelist_names |= set(whitelist_file.read().strip().split('\n'))
+        whitelist_file.close()
+
     self.res = grd_reader.Parse(opts.input, first_id_filename=first_id_filename,
                                 debug=opts.extra_verbose, defines=self.defines)
     self.res.RunGatherers(recursive = True)
@@ -122,6 +137,28 @@ are exported to translation interchange files (e.g. XMB files), etc.
     # output nodes in the file.
     self.scons_targets = None
 
+    # The set of names that are whitelisted to actually be included in the
+    # output.
+    self.whitelist_names = None
+
+
+  # static method
+  def AddWhitelistTags(start_node, whitelist_names):
+    # Walk the tree of nodes added attributes for the nodes that shouldn't
+    # be written into the target files (skip markers).
+    from grit.node import include
+    from grit.node import message
+    for node in start_node.inorder():
+      # Same trick data_pack.py uses to see what nodes actually result in
+      # real items.
+      if (isinstance(node, include.IncludeNode) or
+          isinstance(node, message.MessageNode)):
+        text_ids = node.GetTextualIds()
+        # Mark the item to be skipped if it wasn't in the whitelist.
+        if text_ids and not text_ids[0] in whitelist_names:
+          node.SetWhitelistMarkedAsSkip(True)
+  AddWhitelistTags = staticmethod(AddWhitelistTags)
+
   # static method
   def ProcessNode(node, output_node, outfile):
     '''Processes a node in-order, calling its formatter before and after
@@ -132,13 +169,20 @@ are exported to translation interchange files (e.g. XMB files), etc.
       output_node: grit.node.io.File
       outfile: open filehandle
     '''
+    # See if the node should be skipped by a whitelist.
+    # Note: Some Format calls have side effects, so Format is always called
+    # and the whitelist is used to only avoid the output.
+    should_write = not node.WhitelistMarkedAsSkip()
+
     base_dir = util.dirname(output_node.GetOutputFilename())
 
     try:
       formatter = node.ItemFormatter(output_node.GetType())
       if formatter:
-        outfile.write(formatter.Format(node, output_node.GetLanguage(),
-                                       begin_item=True, output_dir=base_dir))
+        formatted = formatter.Format(node, output_node.GetLanguage(),
+                                     begin_item=True, output_dir=base_dir)
+        if should_write:
+          outfile.write(formatted)
     except:
       print u"Error processing node %s" % unicode(node)
       raise
@@ -148,8 +192,10 @@ are exported to translation interchange files (e.g. XMB files), etc.
 
     try:
       if formatter:
-        outfile.write(formatter.Format(node, output_node.GetLanguage(),
-                                       begin_item=False, output_dir=base_dir))
+        formatted = formatter.Format(node, output_node.GetLanguage(),
+                                     begin_item=False, output_dir=base_dir)
+        if should_write:
+          outfile.write(formatted)
     except:
       print u"Error processing node %s" % unicode(node)
       raise
@@ -172,6 +218,11 @@ are exported to translation interchange files (e.g. XMB files), etc.
         output.output_filename = os.path.abspath(os.path.join(
           self.output_directory, output.GetFilename()))
 
+    # If there are whitelisted names, tag the tree once up front, this way
+    # while looping through the actual output, it is just an attribute check.
+    if self.whitelist_names:
+      self.AddWhitelistTags(self.res, self.whitelist_names)
+
     for output in self.res.GetOutputFiles():
       self.VerboseOut('Creating %s...' % output.GetFilename())
 
diff --git a/tools/grit/grit_info.py b/tools/grit/grit_info.py
index 9bb7c430..700dba3 100755
--- a/tools/grit/grit_info.py
+++ b/tools/grit/grit_info.py
@@ -1,5 +1,5 @@
 #!/usr/bin/python
-# Copyright (c) 2006-2008 The Chromium Authors. All rights reserved.
+# 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.
 
@@ -86,6 +86,7 @@ def main(argv):
   # grit build also supports '-E KEY=VALUE', support that to share command
   # line flags.
   parser.add_option("-E", action="append", dest="build_env", default=[])
+  parser.add_option("-w", action="append", dest="whitelist_files", default=[])
 
   options, args = parser.parse_args()
 
@@ -102,6 +103,8 @@ def main(argv):
       inputs = Inputs(filename, defines)
       # Include grd file as second input (works around gyp expecting it).
       inputs = [inputs[0], filename] + inputs[1:]
+      if options.whitelist_files:
+        inputs.extend(options.whitelist_files)
       print '\n'.join(inputs)
   elif options.outputs:
     if len(args) < 2: