Lib/whichdb.py

   1 # !/usr/bin/env python
   2 """Guess which db package to use to open a db file."""
   3
   4 import os
   5 import struct
   6 import sys
   7
   8 try:
   9     import dbm
  10     _dbmerror = dbm.error
  11 except ImportError:
  12     dbm = None
  13     # just some sort of valid exception which might be raised in the
  14     # dbm test
  15     _dbmerror = IOError
  16
  17 def whichdb(filename):
  18     """Guess which db package to use to open a db file.
  19
  20     Return values:
  21
  22     - None if the database file can't be read;
  23     - empty string if the file can be read but can't be recognized
  24     - the module name (e.g. "dbm" or "gdbm") if recognized.
  25
  26     Importing the given module may still fail, and opening the
  27     database using that module may still fail.
  28     """
  29
  30     # Check for dbm first -- this has a .pag and a .dir file
  31     try:
  32         f = open(filename + os.extsep + "pag", "rb")
  33         f.close()
  34         # dbm linked with gdbm on OS/2 doesn't have .dir file
  35         if not (dbm.library == "GNU gdbm" and sys.platform == "os2emx"):
  36             f = open(filename + os.extsep + "dir", "rb")
  37             f.close()
  38         return "dbm"
  39     except IOError:
  40         # some dbm emulations based on Berkeley DB generate a .db file
  41         # some do not, but they should be caught by the dbhash checks
  42         try:
  43             f = open(filename + os.extsep + "db", "rb")
  44             f.close()
  45             # guarantee we can actually open the file using dbm
  46             # kind of overkill, but since we are dealing with emulations
  47             # it seems like a prudent step
  48             if dbm is not None:
  49                 d = dbm.open(filename)
  50                 d.close()
  51                 return "dbm"
  52         except (IOError, _dbmerror):
  53             pass
  54
  55     # Check for dumbdbm next -- this has a .dir and a .dat file
  56     try:
  57         # First check for presence of files
  58         os.stat(filename + os.extsep + "dat")
  59         size = os.stat(filename + os.extsep + "dir").st_size
  60         # dumbdbm files with no keys are empty
  61         if size == 0:
  62             return "dumbdbm"
  63         f = open(filename + os.extsep + "dir", "rb")
  64         try:
  65             if f.read(1) in ("'", '"'):
  66                 return "dumbdbm"
  67         finally:
  68             f.close()
  69     except (OSError, IOError):
  70         pass
  71
  72     # See if the file exists, return None if not
  73     try:
  74         f = open(filename, "rb")
  75     except IOError:
  76         return None
  77
  78     # Read the start of the file -- the magic number
  79     s16 = f.read(16)
  80     f.close()
  81     s = s16[0:4]
  82
  83     # Return "" if not at least 4 bytes
  84     if len(s) != 4:
  85         return ""
  86
  87     # Convert to 4-byte int in native byte order -- return "" if impossible
  88     try:
  89         (magic,) = struct.unpack("=l", s)
  90     except struct.error:
  91         return ""
  92
  93     # Check for GNU dbm
  94     if magic == 0x13579ace:
  95         return "gdbm"
  96
  97     # Check for old Berkeley db hash file format v2
  98     if magic in (0x00061561, 0x61150600):
  99         return "bsddb185"
 100
 101     # Later versions of Berkeley db hash file have a 12-byte pad in
 102     # front of the file type
 103     try:
 104         (magic,) = struct.unpack("=l", s16[-4:])
 105     except struct.error:
 106         return ""
 107
 108     # Check for BSD hash
 109     if magic in (0x00061561, 0x61150600):
 110         return "dbhash"
 111
 112     # Unknown
 113     return ""
 114
 115 if __name__ == "__main__":
 116     for filename in sys.argv[1:]:
 117         print whichdb(filename) or "UNKNOWN", filename