Created a tool to remove duplicate includes between h and cc files.

tools/remove_duplicate_includes.py

+#!/usr/bin/env python
+# 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.
+
+"""This script will search through the target folder specified and try to find
+duplicate includes from h and cc files, and remove them from the cc files.
+
+Usage remove_duplicate_includes.py --dry-run components/foo components/bar
+"""
+
+import argparse;
+import collections;
+import os;
+import re;
+
+parser = argparse.ArgumentParser()
+parser.add_argument('--dry-run', action='store_true',
+  help='Does not actually remove lines when specified.')
+parser.add_argument('--span-dirs', action='store_true',
+  help='Mapping between h and cc files will not be limited to same folders.')
+parser.add_argument('targets', nargs='+',
+  help="Folders to search for duplicate includes in.")
+args = parser.parse_args()
+
+# This could be generlized if desired, and moved to command line arguments.

[maxbogue, 2016/09/30 17:40:31]

generalized

[skym, 2016/09/30 19:18:12]

Done.

+H_FILE_SUFFIX = ".h"
+CC_FILE_SUFFIX = ".cc"
+
+def HasSuffix(file_name, suffix):
+  return os.path.splitext(file_name)[1] == suffix
+
+def IsEmpty(line):
+  return not line.strip()
+
+# The \s should allow us to ignore any whitespace and only focus on the group
+# captured when comparing between files.
+regex = re.compile('^\s*#include\s+(.*?)\s*$')
+
+# The key here depends on the span-dirs flag, if specified then it will only be

[maxbogue, 2016/09/30 17:40:31]

"A map of header files to the includes they contain. The key is the full path of
the header file unless the span-dirs flag is set, in which case will only be the
file name (to allow mapping between files not in the same folder)."

- The purpose of the variable first
- The default key case (no flag) second
- The abnormal key case (flag) third

[skym, 2016/09/30 19:18:12]

Done.

+# the file name, and this will allows mapping between files not in the same
+# folder. If this flag not present then full path is used.
+h_file_to_include_set = collections.defaultdict(set)
+
+# Key is always the full path to the cc file.

[maxbogue, 2016/09/30 17:40:31]

Explain the purpose of this variable. I think it collects all the cc files
during the traversal so you can go through them after the headers have been
processed?

[skym, 2016/09/30 19:18:12]

Done.

+cc_file_path_set = set()
+
+for relative_root in args.targets:
+  absolute_root = os.path.join(os.getcwd(), relative_root)
+  for (dir_path, dir_name_list, file_name_list) in os.walk(absolute_root):

[maxbogue, 2016/09/30 17:40:31]

I haven't worked with python in a while but I'm pretty sure you don't need the
parens around the tuple.

[skym, 2016/09/30 19:18:12]

Done.

+    for file_name in file_name_list:
+      file_path = os.path.join(dir_path, file_name)
+      if HasSuffix(file_name, H_FILE_SUFFIX):
+        # Can be either name or path depending on flag.
+        file_key = file_name if args.span_dirs else file_path
+        with open(file_path) as file_handle:
+          for line in file_handle:
+            match = regex.search(line)
+            if match:
+              h_file_to_include_set[file_key].add(match.group(1))
+      elif HasSuffix(file_name, CC_FILE_SUFFIX):
+        cc_file_path_set.add(file_path)
+
+for cc_file_path in cc_file_path_set:
+  # The lookup must match index method when adding h files, depending on flag.
+  cc_file_name = os.path.basename(cc_file_path)
+  cc_file_key = cc_file_name if args.span_dirs else cc_file_path
+  h_file_key = os.path.splitext(cc_file_key)[0] + H_FILE_SUFFIX

[maxbogue, 2016/09/30 17:40:31]

You could strip "_unittest" from the end here if it exists and suddenly this
would catch duplicate includes in unittest files as well right?

[skym, 2016/09/30 19:18:13]

Oooooh, actually maybe I should be looking at the very first include in the .cc
file and use that to calculate its appropriate .h file instead. And if it
doesn't conform to the name of file.cc/file.h/file_unittest.cc then throw out a
warning.

+
+  if h_file_key in h_file_to_include_set:
+    include_set = h_file_to_include_set[h_file_key]
+
+    # Read out all the data and reset file position to start overwriting.
+    file_handle = open(cc_file_path, "r" if args.dry_run else "r+")
+    data = file_handle.readlines()

[maxbogue, 2016/09/30 17:40:31]

I'd probably just call this "lines" tbh

[skym, 2016/09/30 19:18:12]

How about line_list?

+    file_handle.seek(0)
+
+    # When a section of includes are completely removed, we want to remove the
+    # trailing empty as well.
+    lastCopiedLineWasEmpty = False
+    lastLineWasOmitted = False
+    for line in data:
+        match = regex.search(line)
+        if match is not None and match.group(1) in include_set:
+          print "Removed " + match.group(1) + " " + cc_file_name
+          lastLineWasOmitted = True
+        elif lastCopiedLineWasEmpty and lastLineWasOmitted and IsEmpty(line):
+          print "Removed empty line " + cc_file_name
+          lastLineWasOmitted = True
+        else:
+          lastCopiedLineWasEmpty = IsEmpty(line)
+          lastLineWasOmitted = False
+          if not args.dry_run:
+            file_handle.write(line)
+    if not args.dry_run:
+      file_handle.truncate()
+    file_handle.close()