mojo/tools/pylib/transitive_hash.py

   1 # Copyright 2014 The Chromium Authors. All rights reserved.
   2 # Use of this source code is governed by a BSD-style license that can be
   3 # found in the LICENSE file.
   4
   5 import logging
   6 import subprocess
   7 import sys
   8
   9 from hashlib import sha256
  10 from os.path import basename, realpath
  11
  12 _logging = logging.getLogger()
  13
  14 # Based on/taken from
  15 #   http://code.activestate.com/recipes/578231-probably-the-fastest-memoization-decorator-in-the-/
  16 # (with cosmetic changes).
  17 def _memoize(f):
  18   """Memoization decorator for a function taking a single argument."""
  19   class Memoize(dict):
  20     def __missing__(self, key):
  21       rv = self[key] = f(key)
  22       return rv
  23   return Memoize().__getitem__
  24
  25 @_memoize
  26 def _file_hash(filename):
  27   """Returns a string representing the hash of the given file."""
  28   _logging.debug("Hashing %s ...", filename)
  29   rv = subprocess.check_output(['sha256sum', '-b', filename]).split(None, 1)[0]
  30   _logging.debug("  => %s", rv)
  31   return rv
  32
  33 @_memoize
  34 def _get_dependencies(filename):
  35   """Returns a list of filenames for files that the given file depends on."""
  36   _logging.debug("Getting dependencies for %s ...", filename)
  37   lines = subprocess.check_output(['ldd', filename]).splitlines()
  38   rv = []
  39   for line in lines:
  40     i = line.find('/')
  41     if i < 0:
  42       _logging.debug("  => no file found in line: %s", line)
  43       continue
  44     rv.append(line[i:].split(None, 1)[0])
  45   _logging.debug("  => %s", rv)
  46   return rv
  47
  48 def transitive_hash(filename):
  49   """Returns a string that represents the "transitive" hash of the given
  50   file. The transitive hash is a hash of the file and all the shared libraries
  51   on which it depends (done in an order-independent way)."""
  52   hashes = set()
  53   to_hash = [filename]
  54   while to_hash:
  55     current_filename = realpath(to_hash.pop())
  56     current_hash = _file_hash(current_filename)
  57     if current_hash in hashes:
  58       _logging.debug("Already seen %s (%s) ...", current_filename, current_hash)
  59       continue
  60     _logging.debug("Haven't seen %s (%s) ...", current_filename, current_hash)
  61     hashes.add(current_hash)
  62     to_hash.extend(_get_dependencies(current_filename))
  63   return sha256('|'.join(sorted(hashes))).hexdigest()
  64
  65 def main(argv):
  66   logging.basicConfig()
  67   # Uncomment to debug:
  68   # _logging.setLevel(logging.DEBUG)
  69
  70   if len(argv) < 2:
  71     print """\
  72 Usage: %s [file] ...
  73
  74 Prints the \"transitive\" hash of each (executable) file. The transitive
  75 hash is a hash of the file and all the shared libraries on which it
  76 depends (done in an order-independent way).""" % basename(argv[0])
  77     return 0
  78
  79   rv = 0
  80   for filename in argv[1:]:
  81     try:
  82       print transitive_hash(filename), filename
  83     except:
  84       print "ERROR", filename
  85       rv = 1
  86   return rv
  87
  88 if __name__ == '__main__':
  89   sys.exit(main(sys.argv))