Remove the timestamps from the zip archives in remove_build_metadata.py

scripts/slave/recipe_modules/isolate/resources/remove_build_metadata.py

 #!/usr/bin/env python
 # Copyright 2014 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.
 """Remove the build metadata embedded in the artifacts of a build."""
 
 import json
 import optparse
 import os
+import shutil
 import subprocess
 import sys
+import tempfile
+import zipfile
 
 
 BASE_DIR = os.path.dirname(os.path.abspath(__file__))
 
 
-def RunZapTimestamp(src_dir, filepath):
+def get_files_to_clean(build_dir, recursive=False):
+  """Get the list of files to clean."""
+  allowed = frozenset(
+      ('', '.apk', '.app', '.dll', '.dylib', '.exe', '.nexe', '.so'))
+  non_x_ok_exts = frozenset(('.apk', '.isolated'))
+  def check(f):
+    if not os.path.isfile(f):

[M-A Ruel, 2014/11/06 18:38:39]

if not os.path.isfile(f) or os.path.basename(f).startswith('.'):

(just to save 2 lines :)

[Sébastien Marchand, 2014/11/06 19:37:31]

Done.

+      return False
+    if os.path.basename(f).startswith('.'):
+      return False
+    ext = os.path.splitext(f)[1]
+    if ext in non_x_ok_exts:

[M-A Ruel, 2014/11/06 18:38:39]

return (ext in non_x_ok_exts) or (ext in allowed and os.access(f, os.X_OK))

to save another 2 lines :)

[Sébastien Marchand, 2014/11/06 19:37:32]

Done.

+      return True
+    return ext in allowed and os.access(f, os.X_OK)
+
+  ret_files = set()
+  for root, dirs, files in os.walk(build_dir):
+    if not recursive:
+      dirs[:] = [d for d in dirs if d.endswith('_apk')]
+    for f in (f for f in files if check(os.path.join(root, f))):
+      ret_files.add(os.path.relpath(os.path.join(root, f), build_dir))
+  return ret_files
+
+
+def run_zap_timestamp(src_dir, filepath):
+  """Run zap_timestamp.exe on a PE binary."""
+  assert sys.platform == 'win32'
   syzygy_dir = os.path.join(
       src_dir, 'third_party', 'syzygy', 'binaries', 'exe')
   zap_timestamp_exe = os.path.join(syzygy_dir, 'zap_timestamp.exe')
   print('Processing: %s' % os.path.basename(filepath))
   proc = subprocess.Popen(
       [zap_timestamp_exe, '--input-image=%s' % filepath, '--overwrite'],
       stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
   log, _ = proc.communicate()
   if proc.returncode != 0:
     print >> sys.stderr, log
   return proc.returncode
 
 
-def RemovePEMetadata(build_dir, src_dir):
+def remove_pe_metadata(filename, src_dir):
   """Remove the build metadata from a PE file."""
-  files = (i for i in os.listdir(build_dir) if i.endswith(('.dll', '.exe')))
+  # Only run zap_timestamp on the PE files for which we have a PDB.
+  ret = 0
+  if os.path.exists(filename + '.pdb'):

[M-A Ruel, 2014/11/06 18:38:39]

is this check still necessary? e.g. zap_timestamp wasn't updated yet.

[Sébastien Marchand, 2014/11/06 19:37:32]

Yes it is, sorry, I'll remove it once we update zap_timestamp.

+    ret = run_zap_timestamp(src_dir, filename)
+  return ret
 
+
+def remove_archive_timestamps(filename):

[M-A Ruel, 2014/11/06 18:38:39]

remove_apk_medata so it fits with remote_pe_metadata

[Sébastien Marchand, 2014/11/06 19:37:31]

Done, but it also works for the zip files :)

+  """Remove the timestamps embedded in a zip archive."""
+  print('Processing: %s' % os.path.basename(filename))
+  zf = zipfile.ZipFile(filename, 'r')

[M-A Ruel, 2014/11/06 18:38:39]

with zipfile.ZipFile(filename, 'r') as zf:

[Sébastien Marchand, 2014/11/06 19:37:32]

Done.

+  # Creates a temporary file.
+  out_file, out_filename = tempfile.mkstemp()

[M-A Ruel, 2014/11/06 18:38:38]

prefix='remote_apk_timestamp'
put in try/finally

[Sébastien Marchand, 2014/11/06 19:37:31]

Done.

+  os.close(out_file)
+  zf_o = zipfile.ZipFile(out_filename, 'w')

[M-A Ruel, 2014/11/06 18:38:39]

with ..

[Sébastien Marchand, 2014/11/06 19:37:32]

Done.

+  # Copy the data from the original file to the new one.
+  for info in zf.infolist():
+    # Overwrite the timestamp with a constant value.
+    info.date_time = (2008, 9, 2, 8, 0 , 0)

[M-A Ruel, 2014/11/06 18:38:39]

I'd prefer 0 than an arbitrary value.

[Sébastien Marchand, 2014/11/06 19:37:31]

Doesn't work with 0... It looks like the minimal accepted value is (1980, 1, 1,
0, 0, 0):
https://github.com/enthought/Python-2.7.3/blob/master/Lib/zipfile.py#L277

+    zf_o.writestr(info, zf.read(info.filename))
+  zf.close()
+  zf_o.close()
+  # Remove the original file and replace it by the modified one.
+  os.remove(filename)
+  shutil.move(out_filename, filename)
+
+
+def remove_metadata(build_dir, src_dir, recursive):
+  """Remove the build metadata from the artifacts of a build."""
   with open(os.path.join(BASE_DIR, 'deterministic_build_blacklist.json')) as f:
     blacklist = frozenset(json.load(f))
+  files = get_files_to_clean(build_dir, recursive) - blacklist
+  failed_files = []
+  ret = 0
+  for f in files:
+    if f.endswith(('.dll', '.exe')):
+      if remove_pe_metadata(os.path.join(build_dir, f), src_dir):
+        ret = 1
+        failed_files.append(f)
+    elif f.endswith('.apk'):
+      remove_archive_timestamps(os.path.join(build_dir, f))
 
-  failed = []
-  for filename in files:
-    # Ignore the blacklisted files.
-    if filename in blacklist:
-      print('Ignored: %s' % filename)
-      continue
-    # Only run zap_timestamp on the PE files for which we have a PDB.
-    if os.path.exists(os.path.join(build_dir, filename + '.pdb')):
-      ret = RunZapTimestamp(src_dir, os.path.join(build_dir, filename))
-      if ret != 0:
-        failed.append(filename)
-
-  if failed:
-    print >> sys.stderr, 'zap_timestamp.exe failed for the following files:'
-    print >> sys.stderr, '\n'.join('  ' + i for i in sorted(failed))
+  if failed_files:
+    print >> sys.stderr, 'Failed for the following files:'
+    print >> sys.stderr, '\n'.join('  ' + i for i in sorted(failed_files))
     return 1
 
-  return 0
+  return ret
 
 
 def main():
   parser = optparse.OptionParser(usage='%prog [options]')
   # TODO(sebmarchand): Add support for reading the list of artifact from a
   # .isolated file.
   parser.add_option('--build-dir', help='The build directory.')
   parser.add_option('--src-dir', help='The source directory.')
+  parser.add_option('-r', '--recursive', action='store_true', default=False,
+                    help='Indicates if the script should be recursive.')
   options, _ = parser.parse_args()
 
   if not options.build_dir:
     parser.error('--build-dir is required')
-  if not options.src_dir:
+
+  if sys.platform == 'win32' and not options.src_dir:
     parser.error('--src-dir is required')
 
-  # There's nothing to do for the non-Windows platform yet.
-  if sys.platform == 'win32':
-    return RemovePEMetadata(options.build_dir, options.src_dir)
+  return remove_metadata(options.build_dir, options.src_dir, options.recursive)
 
 
 if __name__ == '__main__':
   sys.exit(main())