etc/iOS/parse_manifest.py

   1 #!/usr/bin/env python
   2 # From http://stackoverflow.com/questions/3085153/how-to-parse-the-manifest-mbdb-file-in-an-ios-4-0-itunes-backup
   3 import sys
   4 import hashlib
   5
   6 mbdx = {}
   7
   8 def getint(data, offset, intsize):
   9     """Retrieve an integer (big-endian) and new offset from the current offset"""
  10     value = 0
  11     while intsize > 0:
  12         value = (value<<8) + ord(data[offset])
  13         offset = offset + 1
  14         intsize = intsize - 1
  15     return value, offset
  16
  17 def getstring(data, offset):
  18     """Retrieve a string and new offset from the current offset into the data"""
  19     if data[offset] == chr(0xFF) and data[offset+1] == chr(0xFF):
  20         return '', offset+2 # Blank string
  21     length, offset = getint(data, offset, 2) # 2-byte length
  22     value = data[offset:offset+length]
  23     return value, (offset + length)
  24
  25 def process_mbdb_file(filename):
  26     mbdb = {} # Map offset of info in this file => file info
  27     data = open(filename).read()
  28     if data[0:4] != "mbdb": raise Exception("This does not look like an MBDB file")
  29     offset = 4
  30     offset = offset + 2 # value x05 x00, not sure what this is
  31     while offset < len(data):
  32         fileinfo = {}
  33         fileinfo['start_offset'] = offset
  34         fileinfo['domain'], offset = getstring(data, offset)
  35         fileinfo['filename'], offset = getstring(data, offset)
  36         fileinfo['linktarget'], offset = getstring(data, offset)
  37         fileinfo['datahash'], offset = getstring(data, offset)
  38         fileinfo['unknown1'], offset = getstring(data, offset)
  39         fileinfo['mode'], offset = getint(data, offset, 2)
  40         fileinfo['unknown2'], offset = getint(data, offset, 4)
  41         fileinfo['unknown3'], offset = getint(data, offset, 4)
  42         fileinfo['userid'], offset = getint(data, offset, 4)
  43         fileinfo['groupid'], offset = getint(data, offset, 4)
  44         fileinfo['mtime'], offset = getint(data, offset, 4)
  45         fileinfo['atime'], offset = getint(data, offset, 4)
  46         fileinfo['ctime'], offset = getint(data, offset, 4)
  47         fileinfo['filelen'], offset = getint(data, offset, 8)
  48         fileinfo['flag'], offset = getint(data, offset, 1)
  49         fileinfo['numprops'], offset = getint(data, offset, 1)
  50         fileinfo['properties'] = {}
  51         for ii in range(fileinfo['numprops']):
  52             propname, offset = getstring(data, offset)
  53             propval, offset = getstring(data, offset)
  54             fileinfo['properties'][propname] = propval
  55         mbdb[fileinfo['start_offset']] = fileinfo
  56         fullpath = fileinfo['domain'] + '-' + fileinfo['filename']
  57         id = hashlib.sha1(fullpath)
  58         mbdx[fileinfo['start_offset']] = id.hexdigest()
  59     return mbdb
  60
  61 def modestr(val):
  62     def mode(val):
  63         if (val & 0x4): r = 'r'
  64         else: r = '-'
  65         if (val & 0x2): w = 'w'
  66         else: w = '-'
  67         if (val & 0x1): x = 'x'
  68         else: x = '-'
  69         return r+w+x
  70     return mode(val>>6) + mode((val>>3)) + mode(val)
  71
  72 def fileinfo_str(f, verbose=False):
  73     # if not verbose: return "(%s)%s::%s" % (f['fileID'], f['domain'], f['filename'])
  74     if not verbose: return "%s => %s (%s)" % (f['fileID'], f['filename'], f['domain'])
  75     if (f['mode'] & 0xE000) == 0xA000: type = 'l' # symlink
  76     elif (f['mode'] & 0xE000) == 0x8000: type = '-' # file
  77     elif (f['mode'] & 0xE000) == 0x4000: type = 'd' # dir
  78     else:
  79         print >> sys.stderr, "Unknown file type %04x for %s" % (f['mode'], fileinfo_str(f, False))
  80         type = '?' # unknown
  81     info = ("%s%s %08x %08x %7d %10d %10d %10d (%s)%s::%s" %
  82             (type, modestr(f['mode']&0x0FFF) , f['userid'], f['groupid'], f['filelen'],
  83              f['mtime'], f['atime'], f['ctime'], f['fileID'], f['domain'], f['filename']))
  84     if type == 'l': info = info + ' -> ' + f['linktarget'] # symlink destination
  85     for name, value in f['properties'].items(): # extra properties
  86         info = info + ' ' + name + '=' + repr(value)
  87     return info
  88
  89 verbose = True
  90 if __name__ == '__main__':
  91     mbdb = process_mbdb_file("Manifest.mbdb")
  92     for offset, fileinfo in mbdb.items():
  93         if offset in mbdx:
  94             fileinfo['fileID'] = mbdx[offset]
  95         else:
  96             fileinfo['fileID'] = "<nofileID>"
  97             print >> sys.stderr, "No fileID found for %s" % fileinfo_str(fileinfo)
  98         print fileinfo_str(fileinfo)