Add script fix_gn_headers.py to fix GN missing headers

build/fix_gn_headers.py

Index: build/fix_gn_headers.py
diff --git a/build/fix_gn_headers.py b/build/fix_gn_headers.py
new file mode 100755
index 0000000000000000000000000000000000000000..14e8ff061badc7e2332e6698abe014c35154a6da
--- /dev/null
+++ b/build/fix_gn_headers.py
@@ -0,0 +1,122 @@
+#!/usr/bin/env python
+# Copyright 2017 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.
+
+"""Fix header files missing in GN.
+
+This script takes the missing header files from check_gn_headers.py, and
+try to fix them by adding them to the GN files.
+Manual cleaning up is likely required afterwards.
+"""
+
+import argparse
+import os
+import re
+import subprocess
+import sys
+
+
+def AddHeadersNextToCC(headers, skip_ambiguous=True):
+  """Add header files next to the corresponding .cc files in GN files.
+
+  When skip_ambiguous is True, skip if multiple .cc files are found.
+  Returns unhandled headers.
+
+  Manual cleaning up is likely required, especially if not skip_ambiguous.
+  """
+  edits = {}
+  unhandled = []
+  for filename in headers:
+    filename = filename.strip()
+    if not (filename.endswith('.h') or filename.endswith('.hh')):

[Nico, 2017/03/28 14:37:55]

Do we have any headers ending in .hh?

[wychen, 2017/03/28 17:52:05]

Yes, we do, and some of them are part of the GN missing headers.

+      continue
+    basename = os.path.basename(filename)
+    print filename
+    cc = r'\b' + os.path.splitext(basename)[0] + r'\.(cc|cpp|mm)\b'
+    p = subprocess.Popen(
+        ['git', 'grep', '-En', cc + '"', '--', '*.gn', '*.gni'],
+        stdout=subprocess.PIPE)
+    out, _ = p.communicate()
+    if p.returncode != 0 or not out:
+      unhandled.append(filename)
+      continue
+
+    if skip_ambiguous and len(out.splitlines()) > 1:
+      print '\n[WARNING] Ambiguous matching for', filename
+      print out
+      continue
+
+    for gnline in out.splitlines():
+      gnfile, linenr, contents = gnline.split(':')
+      linenr = int(linenr)
+      new = re.sub(cc, basename, contents)
+      lines = open(gnfile).read().splitlines()
+      if lines[linenr] == new:
+        continue
+      if lines[linenr - 2] == new:

[Nico, 2017/03/28 14:37:55]

why this?

[wychen, 2017/03/28 17:52:05]

This and the line above is for deduplication.
lines[linenr] is the line after the match, and lines[linenr-2] is before.
Depending on whether that match is a cc or mm, the .h could be before or after.

+        continue
+      print '  ', gnfile, linenr, new
+      edits.setdefault(gnfile, {})[linenr] = new
+
+  for gnfile in edits:
+    lines = open(gnfile).read().splitlines()
+    for l in reversed(sorted(edits[gnfile].keys())):

[Nico, 2017/03/28 14:37:55]

Ah, that was the bug! I had assumed the keys were ordered, but you're right,
they aren't. Write this as `sorted(edits[gnfile].keys(), reverse=True)` though.

[wychen, 2017/03/28 17:52:05]

Done.

+      lines.insert(l, edits[gnfile][l])
+    open(gnfile, 'w').write('\n'.join(lines) + '\n')
+
+  return unhandled
+
+
+def AddHeadersToSources(headers, skip_ambiguous=True):
+  """Add header files to the sources list in the closest GN file.

[Nico, 2017/03/28 14:37:55]

"to the _first_ sources list"

Mention somehow that this usually does the wrong thing for _test files if the
test and the main target are in the same .gn file.

[wychen, 2017/03/28 17:52:05]

Done.

+
+  When skip_ambiguous is True, skip if multiple sources arrays are found.
+
+  "git cl format" afterwards is required. Manually cleaning up duplicated items
+  is likely required.
+  """
+  for filename in headers:
+    filename = filename.strip()
+    print filename
+    dirname = os.path.dirname(filename)
+    while not os.path.exists(os.path.join(dirname, 'BUILD.gn')):
+      dirname = os.path.dirname(dirname)
+    rel = filename[len(dirname) + 1:]
+    gnfile = os.path.join(dirname, 'BUILD.gn')
+
+    lines = open(gnfile).read().splitlines()
+    matched = [i for i, l in enumerate(lines) if ' sources = [' in l]
+    if skip_ambiguous and len(matched) > 1:
+      print '[WARNING] Multiple sources in', gnfile
+      continue
+
+    if len(matched) < 1:
+      continue
+    print '  ', gnfile, rel
+    index = matched[0]
+    lines.insert(index + 1, '"%s",' % rel)
+    open(gnfile, 'w').write('\n'.join(lines) + '\n')
+
+
+def main():
+  parser = argparse.ArgumentParser()
+  parser.add_argument('--input', required=True,
+                      help="missing headers, output of check_gn_headers.py")

[Nico, 2017/03/28 14:37:55]

nit: I'd make this one a positional argument

[wychen, 2017/03/28 17:52:05]

Done.

+  parser.add_argument('--prefix',
+                      help="only handle path name with this prefix")
+  parser.add_argument('args', nargs=argparse.REMAINDER)
+
+  args, _extras = parser.parse_known_args()
+
+  headers = open(args.input).readlines()
+
+  if args.prefix:
+    headers = [i for i in headers if i.startswith(args.prefix)]
+
+  unhandled = AddHeadersNextToCC(headers)
+  AddHeadersToSources(unhandled)
+
+
+if __name__ == '__main__':
+  sys.exit(main())