source: releases/1.0.0/python/pyregfi/__init__.py @ 294

Last change on this file since 294 was 262, checked in by tim, 13 years ago

changed regfi_conv_charset to handle memory allocation
tweaked test cases
corrected some documentation

File size: 36.1 KB
Line 
1#!/usr/bin/env python
2
3# Copyright (C) 2010-2011 Timothy D. Morgan
4#
5# This program is free software; you can redistribute it and/or modify
6# it under the terms of the GNU General Public License as published by
7# the Free Software Foundation; version 3 of the License.
8#
9# This program is distributed in the hope that it will be useful,
10# but WITHOUT ANY WARRANTY; without even the implied warranty of
11# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
12# GNU General Public License for more details.
13#
14# You should have received a copy of the GNU General Public License
15# along with this program; if not, write to the Free Software
16# Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
17#
18# $Id: $
19
20## @package pyregfi
21# Python interface to the regfi library.
22#
23
24## @mainpage API Documentation
25#
26# The pyregfi module provides a Python interface to the @ref regfi Windows
27# registry library. 
28#
29# The library operates on registry hives, each of which is contained within a
30# single file.  The quickest way to get started, is to use the @ref openHive()
31# function to obtain a Hive object.  For example:
32# @code
33# >>> import pyregfi
34# >>> myHive = pyregfi.openHive('/mnt/win/c/WINDOWS/system32/config/system')
35# @endcode
36#
37# Using this Hive object, one can begin investigating what top-level keys
38# exist by starting with the root Key attribute:
39# @code
40# >>> for key in myHive.root.subkeys:
41# ...   print(key.name)
42# ControlSet001
43# ControlSet003
44# LastKnownGoodRecovery
45# MountedDevices
46# Select
47# Setup
48# WPA
49# @endcode
50#
51# From there, accessing subkeys and values by name is a simple matter of:
52# @code
53# >>> myKey = myHive.root.subkeys['Select']
54# >>> myValue = myKey.values['Current']
55# @endcode
56#
57# The data associated with a Value can be obtained through the fetch_data()
58# method:
59# @code
60# >>> print(myValue.fetch_data())
61# 1
62# @endcode
63#
64# While useful for simple exercises, using the subkeys object for deeply nested
65# paths is not efficient and doesn't make for particularly attractive code. 
66# Instead, a special-purpose HiveIterator class is provided for simplicity of
67# use and fast access to specific known paths:
68# @code
69# >>> myIter = pyregfi.HiveIterator(myHive)
70# >>> myIter.descend(['ControlSet001','Control','NetworkProvider','HwOrder'])
71# >>> myKey = myIter.current_key()
72# >>> print(myKey.values['ProviderOrder'].fetch_data())
73# RDPNP,LanmanWorkstation,WebClient
74# @endcode
75#
76# The first two lines above can be simplified in some "syntactic sugar" provided
77# by the Hive.subtree() method.  Also, as one might expect, the HiveIterator
78# also acts as an iterator, producing keys in a depth-first order.
79# For instance, to traverse all keys under the ControlSet003\\Services key,
80# printing their names as we go, we could do:
81# @code
82# >>> for key in Hive.subtree(['ControlSet003','Services']):
83# >>>   print(key.name)
84# Services
85# Abiosdsk
86# abp480n5
87# Parameters
88# PnpInterface
89# ACPI
90# [...]
91# @endcode
92#
93# Note that "Services" was printed first, since the subtree is traversed as a
94# "preordering depth-first" search starting with the HiveIterator's current_key(). 
95# As one might expect, traversals of subtrees stops when all elements in a
96# specific subtree (and none outside of it) have been traversed.
97#
98# For more information, peruse the various attributes and methods available on
99# the Hive, HiveIterator, Key, Value, and Security classes.
100#
101# @note @ref regfi is a read-only library by design and there
102# are no plans to implement write support.
103#
104# @note At present, pyregfi has been tested with Python versions 2.6 and 3.1
105#
106# @note Developers strive to make pyregfi thread-safe.
107#
108import sys
109import time
110import ctypes
111import ctypes.util
112import threading
113from pyregfi.structures import *
114
115
116## An enumeration of registry Value data types
117#
118# @note This is a static class, there is no need to instantiate it.
119#       Just access its attributes directly as DATA_TYPES.SZ, etc
120class DATA_TYPES(object):
121    ## None / Unknown
122    NONE                       =  0
123    ## String
124    SZ                         =  1
125    ## String with %...% expansions
126    EXPAND_SZ                  =  2
127    ## Binary buffer
128    BINARY                     =  3
129    ## 32 bit integer (little endian)
130    DWORD                      =  4 # DWORD, little endian
131    ## 32 bit integer (little endian)
132    DWORD_LE                   =  4
133    ## 32 bit integer (big endian)
134    DWORD_BE                   =  5 # DWORD, big endian
135    ## Symbolic link
136    LINK                       =  6
137    ## List of strings
138    MULTI_SZ                   =  7
139    ## Unknown structure
140    RESOURCE_LIST              =  8
141    ## Unknown structure
142    FULL_RESOURCE_DESCRIPTOR   =  9
143    ## Unknown structure
144    RESOURCE_REQUIREMENTS_LIST = 10
145    ## 64 bit integer
146    QWORD                      = 11 # 64-bit little endian
147
148
149## An enumeration of log message types
150#
151# @note This is a static class, there is no need to instantiate it.
152#       Just access its attributes directly as LOG_TYPES.INFO, etc
153class LOG_TYPES(object):
154    ## Informational messages, useful in debugging
155    INFO  =  0x01
156    ## Non-critical problems in structure parsing or intepretation
157    WARN  =  0x04
158    ## Major failures
159    ERROR =  0x10
160
161
162def _buffer2bytearray(char_pointer, length):
163    if length == 0 or char_pointer == None:
164        return None
165   
166    ret_val = bytearray(length)
167    for i in range(0,length):
168        ret_val[i] = char_pointer[i][0]
169
170    return ret_val
171
172
173def _strlist2charss(str_list):
174    ret_val = []
175    for s in str_list:
176        ret_val.append(s.encode('utf-8', 'replace'))
177
178    ret_val = (ctypes.c_char_p*(len(str_list)+1))(*ret_val)
179    # Terminate the char** with a NULL pointer
180    ret_val[-1] = 0
181
182    return ret_val
183
184
185def _charss2strlist(chars_pointer):
186    ret_val = []
187    i = 0
188    s = chars_pointer[i]
189    while s:
190        ret_val.append(s.decode('utf-8', 'replace'))
191        i += 1
192        s = chars_pointer[i]
193
194    return ret_val
195
196
197
198## Returns the (py)regfi library version
199#
200# @return A string indicating the version
201def getVersion():
202    return regfi.regfi_version()
203
204
205## Retrieves messages produced by regfi during parsing and interpretation
206#
207# The regfi C library may generate log messages stored in a special thread-safe
208# global data structure.  These messages should be retrieved periodically or
209# after each major operation by callers to determine if any errors or warnings
210# should be reported to the user.  Failure to retrieve these could result in
211# excessive memory consumption.
212def getLogMessages():
213    msgs = regfi.regfi_log_get_str()
214    if not msgs:
215        return ''
216    return msgs.decode('utf-8')
217
218
219## Sets the types of log messages to record
220#
221# @param log_types A sequence of message types that regfi should generate.
222#                  Message types can be found in the LOG_TYPES enumeration.
223#
224# @return True on success, False on failure.  Failures are rare, but could
225#         indicate that global logging is not operating as expected.
226#
227# Example:
228# @code
229# setLogMask((LOG_TYPES.ERROR, LOG_TYPES.WARN, LOG_TYPES.INFO))
230# @endcode
231#
232# The message mask is a global (all hives, iterators), thread-specific value.
233# For more information, see @ref regfi_log_set_mask.
234#
235def setLogMask(log_types):
236    mask = 0
237    for m in log_types:
238        mask |= m
239    return regfi.regfi_log_set_mask(mask)
240
241
242## Opens a file as a registry hive
243#
244# @param path The file path of a hive, as one would provide to the
245#             open() built-in
246#
247# @return A new Hive instance
248def openHive(path):
249    fh = open(path, 'rb')
250    return Hive(fh)
251
252
253## Abstract class for most objects returned by the library
254class _StructureWrapper(object):
255    _hive = None
256    _base = None
257
258    def __init__(self, hive, base):
259        if not hive:
260            raise Exception("Could not create _StructureWrapper,"
261                            + " hive is NULL.  Current log:\n"
262                            + getLogMessages())
263        if not base:
264            raise Exception("Could not create _StructureWrapper,"
265                            + " base is NULL.  Current log:\n"
266                            + getLogMessages())
267        self._hive = hive
268        self._base = base
269
270
271    # Memory management for most regfi structures is taken care of here
272    def __del__(self):
273        if self._base:
274            regfi.regfi_free_record(self._hive.file, self._base)
275
276
277    # Any attribute requests not explicitly defined in subclasses gets passed
278    # to the equivalent REGFI_* structure defined in structures.py
279    def __getattr__(self, name):
280        return getattr(self._base.contents, name)
281
282   
283    ## Test for equality
284    #
285    # Records returned by pyregfi may be compared with one another.  For example:
286    # @code
287    #  >>> key2 = key1.subkeys['child']
288    #  >>> key1 == key2
289    #  False
290    #  >>> key1 != key2
291    #  True
292    #  >>> key1 == key2.get_parent()
293    #  True
294    # @endcode
295    def __eq__(self, other):
296        return (type(self) == type(other)) and (self.offset == other.offset)
297
298
299    def __ne__(self, other):
300        return (not self.__eq__(other))
301
302
303class Key():
304    pass
305
306
307class Value():
308    pass
309
310
311
312## Represents a registry SK record which contains a security descriptor
313#
314class Security(_StructureWrapper):
315    ## Number of registry Keys referencing this SK record
316    ref_count = 1
317
318    ## The absolute file offset of the SK record's cell in the Hive file
319    offset = 0xCAFEBABE
320
321    ## The @ref winsec.SecurityDescriptor for this SK record
322    descriptor = object()
323
324    def __init__(self, hive, base):
325        super(Security, self).__init__(hive, base)
326        # XXX: add checks for NULL pointers
327        self.descriptor = winsec.SecurityDescriptor(base.contents.sec_desc.contents)
328
329    ## Loads the "next" Security record in the hive
330    #
331    # @note
332    # SK records are included in a circular, doubly-linked list.
333    # To iterate over all SK records, be sure to check for the repetition of
334    # the SK record you started with to determine when all have been traversed.
335    def next_security(self):
336        return Security(self._hive,
337                        regfi.regfi_next_sk(self._hive.file, self._base))
338
339    ## Loads the "previous" Security record in the hive
340    #
341    # @note
342    # SK records are included in a circular, doubly-linked list.
343    # To iterate over all SK records, be sure to check for the repetition of
344    # the SK record you started with to determine when all have been traversed.
345    def prev_security(self):
346        return Security(self._hive,
347                        regfi.regfi_prev_sk(self._hive.file, self._base))
348
349
350## Abstract class for ValueList and SubkeyList
351class _GenericList(object):
352    # XXX: consider implementing keys(), values(), items() and other dictionary methods
353    _hive = None
354    _key_base = None
355    _length = None
356    _current = None
357
358    # implementation-specific functions for SubkeyList and ValueList
359    _fetch_num = None
360    _find_element = None
361    _get_element = None
362    _constructor = None
363
364    def __init__(self, key):
365        if not key:
366            raise Exception("Could not create _GenericList; key is NULL."
367                            + "Current log:\n" + getLogMessages())
368
369        base = regfi.regfi_reference_record(key._hive.file, key._base)
370        if not base:
371            raise Exception("Could not create _GenericList; memory error."
372                            + "Current log:\n" + getLogMessages())
373        self._key_base = cast(base, type(key._base))
374        self._length = self._fetch_num(self._key_base)
375        self._hive = key._hive
376
377   
378    def __del__(self):
379        regfi.regfi_free_record(self._hive.file, self._key_base)
380
381
382    ## Length of list
383    def __len__(self):
384        return self._length
385
386
387    ## Retrieves a list element by name
388    #
389    # @param name The name of the subkey or value desired. 
390    #             This is case-insensitive.
391    #
392    # @note The registry format does not inherently prevent multiple
393    #       subkeys or values from having the same name, having a key
394    #       and a value with the same name, or having the same name in
395    #       different cases that could both match.
396    #       This interface simply returns the first match in the list.
397    #       Lookups using this method could also fail due to incorrectly
398    #       encoded strings stored as names.
399    #       To identify any duplicates or elements with malformed names,
400    #       use the iterator interface to check every list element.
401    #
402    # @return the first element whose name matches, or None if the element
403    #         could not be found
404    def __getitem__(self, name):
405        # XXX: Consider interpreting integer names as offsets in the underlying list
406        index = ctypes.c_uint32()
407        if isinstance(name, str):
408            name = name.encode('utf-8')
409
410        if name != None:
411            name = create_string_buffer(bytes(name))
412
413        if self._find_element(self._hive.file, self._key_base, 
414                              name, byref(index)):
415            return self._constructor(self._hive,
416                                     self._get_element(self._hive.file,
417                                                       self._key_base,
418                                                       index))
419        raise KeyError('')
420
421
422    ## Fetches the requested element by name, or the default value if the lookup
423    #  fails.
424    #
425    def get(self, name, default):
426        try:
427            return self[name]
428        except KeyError:
429            return default
430   
431    def __iter__(self):
432        self._current = 0
433        return self
434   
435    def __next__(self):
436        if self._current >= self._length:
437            raise StopIteration('')
438
439        elem = self._get_element(self._hive.file, self._key_base,
440                                 ctypes.c_uint32(self._current))
441        self._current += 1
442        return self._constructor(self._hive, elem)
443   
444    # For Python 2.x
445    next = __next__
446
447
448## The list of subkeys associated with a Key
449#
450# This attribute is both iterable:
451# @code
452#   for k in myKey.subkeys:
453#     ...
454# @endcode
455# and accessible as a dictionary:
456# @code
457#   mySubkey = myKey.subkeys["keyName"]
458# @endcode
459#
460# You may also request the len() of a subkeys list.
461# However keys(), values(), items() and similar methods are not currently
462# implemented.
463class SubkeyList(_GenericList):
464    _fetch_num = regfi.regfi_fetch_num_subkeys
465    _find_element = regfi.regfi_find_subkey
466    _get_element = regfi.regfi_get_subkey
467
468
469## The list of values associated with a Key
470#
471# This attribute is both iterable:
472# @code
473#   for v in myKey.values:
474#     ...
475# @endcode
476# and accessible as a dictionary:
477# @code
478#   myValue = myKey.values["valueName"]
479# @endcode
480#
481# You may also request the len() of a values list.
482# However keys(), values(), items() and similar methods are not currently
483# implemented.
484class ValueList(_GenericList):
485    _fetch_num = regfi.regfi_fetch_num_values
486    _find_element = regfi.regfi_find_value
487    _get_element = regfi.regfi_get_value
488
489
490## Registry key
491# These represent registry keys (@ref REGFI_NK records) and provide
492# access to their subkeys, values, and other metadata.
493#
494# @note Key instances may provide access to more attributes than are
495#       documented here.  However, undocumented attributes may change over time
496#       and are not officially supported.  If you need access to an attribute
497#       not shown here, see @ref pyregfi.structures.
498class Key(_StructureWrapper):
499    ## A @ref ValueList object representing the list of Values
500    #  stored on this Key
501    values = None
502
503    ## A @ref SubkeyList object representing the list of subkeys
504    #  stored on this Key
505    subkeys = None
506
507    ## The raw Key name as an uninterpreted bytearray
508    name_raw = (b"...")
509   
510    ## The name of the Key as a (unicode) string
511    name = "..."
512   
513    ## The absolute file offset of the Key record's cell in the Hive file
514    offset = 0xCAFEBABE
515
516    ## This Key's last modified time represented as the number of seconds
517    #  since the UNIX epoch in UTC; similar to what time.time() returns
518    modified = 1300000000.123456
519
520    ## The NK record's flags field
521    flags = 0x10110001
522
523    def __init__(self, hive, base):
524        super(Key, self).__init__(hive, base)
525        self.values = ValueList(self)
526        self.subkeys = SubkeyList(self)
527
528    def __getattr__(self, name):
529        if name == "name":
530            ret_val = super(Key, self).__getattr__(name)
531
532            if not ret_val:
533                ret_val = self.name_raw
534            else:
535                ret_val = ret_val.decode('utf-8', 'replace')
536               
537        elif name == "name_raw":
538            ret_val = super(Key, self).__getattr__(name)
539            length = super(Key, self).__getattr__('name_length')
540            ret_val = _buffer2bytearray(ret_val, length)
541       
542        elif name == "modified":
543            ret_val = regfi.regfi_nt2unix_time(self._base.contents.mtime)
544
545        else:
546            ret_val = super(Key, self).__getattr__(name)
547
548        return ret_val
549
550
551    ## Retrieves the Security properties for this key
552    def fetch_security(self):
553        return Security(self._hive,
554                        regfi.regfi_fetch_sk(self._hive.file, self._base))
555
556
557    ## Retrieves the class name for this key
558    #
559    # Class names are typically stored as UTF-16LE strings, so these are decoded
560    # into proper python (unicode) strings.  However, if this fails, a bytearray
561    # is instead returned containing the raw buffer stored for the class name.
562    #
563    # @return The class name as a string or bytearray.  None if a class name
564    #         doesn't exist or an unrecoverable error occurred during retrieval.
565    def fetch_classname(self):
566        ret_val = None
567        cn_p = regfi.regfi_fetch_classname(self._hive.file, self._base)
568        if cn_p:
569            cn_struct = cn_p.contents
570            if cn_struct.interpreted:
571                ret_val = cn_struct.interpreted.decode('utf-8', 'replace')
572            else:
573                ret_val = _buffer2bytearray(cn_struct.raw,
574                                            cn_struct.size)
575            regfi.regfi_free_record(self._hive.file, cn_p)
576
577        return ret_val
578
579
580    ## Retrieves this key's parent key
581    #
582    # @return The parent's Key instance or None if current key is root
583    #         (or an error occured)
584    def get_parent(self):
585        if self.is_root():
586            return None
587        parent_base = regfi.regfi_get_parentkey(self._hive.file, self._base)
588        if parent_base:
589            return Key(self._hive, parent_base)
590        return None
591
592
593    ## Checks to see if this Key is the root of its Hive
594    #
595    #  @return True if it is, False otherwise
596    def is_root(self):
597        return (self._hive.root == self)
598
599
600## Registry value (metadata)
601#
602# These represent registry values (@ref REGFI_VK records) and provide
603# access to their associated data.
604#
605# @note Value instances may provide access to more attributes than are
606#       documented here.  However, undocumented attributes may change over time
607#       and are not officially supported.  If you need access to an attribute
608#       not shown here, see @ref pyregfi.structures.
609class Value(_StructureWrapper):
610    ## The raw Value name as an uninterpreted bytearray
611    name_raw = (b"...")
612   
613    ## The name of the Value as a (unicode) string
614    name = "..."
615   
616    ## The absolute file offset of the Value record's cell in the Hive file
617    offset = 0xCAFEBABE
618
619    ## The length of data advertised in the VK record
620    data_size = 0xCAFEBABE
621
622    ## An integer which represents the data type for this Value's data
623    # Typically this value is one of 12 types defined in @ref DATA_TYPES,
624    # but in some cases (the SAM hive) it may be used for other purposes
625    type = DATA_TYPES.NONE
626
627    ## The VK record's flags field
628    flags = 0x10110001
629
630    ## Retrieves the Value's data according to advertised type
631    #
632    # Data is loaded from its cell(s) and then interpreted based on the data
633    # type recorded in the Value.  It is not uncommon for data to be stored with
634    # the wrong type or even with invalid types.  If you have difficulty
635    # obtaining desired data here, use @ref fetch_raw_data().
636    #
637    # @return The interpreted representation of the data as one of several
638    #         possible Python types, as listed below.  None if any failure
639    #         occurred during extraction or conversion.
640    #
641    # @retval string for SZ, EXPAND_SZ, and LINK
642    # @retval int for DWORD, DWORD_BE, and QWORD
643    # @retval list(string) for MULTI_SZ
644    # @retval bytearray for NONE, BINARY, RESOURCE_LIST,
645    #         FULL_RESOURCE_DESCRIPTOR, and RESOURCE_REQUIREMENTS_LIST
646    #
647    def fetch_data(self):
648        ret_val = None
649        data_p = regfi.regfi_fetch_data(self._hive.file, self._base)
650        if not data_p:
651            return None
652        data_struct = data_p.contents
653
654        if data_struct.interpreted_size == 0:
655            ret_val = None
656        elif data_struct.type in (DATA_TYPES.SZ, DATA_TYPES.EXPAND_SZ, DATA_TYPES.LINK):
657            # Unicode strings
658            ret_val = data_struct.interpreted.string.decode('utf-8', 'replace')
659        elif data_struct.type in (DATA_TYPES.DWORD, DATA_TYPES.DWORD_BE):
660            # 32 bit integers
661            ret_val = data_struct.interpreted.dword
662        elif data_struct.type == DATA_TYPES.QWORD:
663            # 64 bit integers
664            ret_val = data_struct.interpreted.qword
665        elif data_struct.type == DATA_TYPES.MULTI_SZ:
666            ret_val = _charss2strlist(data_struct.interpreted.multiple_string)
667        elif data_struct.type in (DATA_TYPES.NONE, DATA_TYPES.RESOURCE_LIST,
668                                  DATA_TYPES.FULL_RESOURCE_DESCRIPTOR,
669                                  DATA_TYPES.RESOURCE_REQUIREMENTS_LIST,
670                                  DATA_TYPES.BINARY):
671            ret_val = _buffer2bytearray(data_struct.interpreted.none,
672                                        data_struct.interpreted_size)
673
674        regfi.regfi_free_record(self._hive.file, data_p)
675        return ret_val
676   
677
678    ## Retrieves raw representation of Value's data
679    #
680    # @return A bytearray containing the data
681    #
682    def fetch_raw_data(self):
683        ret_val = None
684        # XXX: should we load the data without interpretation instead?
685        data_p = regfi.regfi_fetch_data(self._hive.file, self._base)
686        if not data_p:
687            return None
688
689        data_struct = data_p.contents
690        ret_val = _buffer2bytearray(data_struct.raw,
691                                    data_struct.size)
692        regfi.regfi_free_record(self._hive.file, data_p)
693        return ret_val
694
695
696    def __getattr__(self, name):
697        ret_val = super(Value, self).__getattr__(name)
698        if name == "name":
699            if not ret_val:
700                ret_val = self.name_raw
701            else:
702                ret_val = ret_val.decode('utf-8', 'replace')
703
704        elif name == "name_raw":
705            length = super(Value, self).__getattr__('name_length')
706            ret_val = _buffer2bytearray(ret_val, length)
707
708        return ret_val
709
710
711# Avoids chicken/egg class definitions.
712# Also makes for convenient code reuse in these lists' parent classes.
713SubkeyList._constructor = Key
714ValueList._constructor = Value
715
716
717
718## Represents a single registry hive (file)
719class Hive():
720    file = None
721    raw_file = None
722    _fh = None
723    #_root = None
724
725
726    ## The root Key of this Hive
727    root = None
728
729    ## This Hives's last modified time represented as the number of seconds
730    #  since the UNIX epoch in UTC; similar to what time.time() returns
731    modified = 1300000000.123456
732
733    ## First sequence number
734    sequence1 = 12345678
735
736    ## Second sequence number
737    sequence2 = 12345678
738
739    ## Major version
740    major_version = 1
741
742    ## Minor version
743    minor_version = 5
744
745    ## Constructor
746    #
747    # Initialize a new Hive based on a Python file object.  To open a file by
748    # path, see @ref openHive.
749    #
750    # @param fh A Python file object.  The constructor first looks for a valid
751    #           fileno attribute on this object and uses it if possible. 
752    #           Otherwise, the seek and read methods are used for file
753    #           access.
754    #
755    # @note Supplied file must be seekable.  Do not perform any operation on
756    #       the provided file object while a Hive is using it.  Do not
757    #       construct multiple Hive instances from the same file object.
758    #       If a file must be accessed by separate code and pyregfi
759    #       simultaneously, use a separate file descriptor.  Hives are
760    #       thread-safe, so multiple threads may use a single Hive object.
761    def __init__(self, fh):
762        # The fileno method may not exist, or it may throw an exception
763        # when called if the file isn't backed with a descriptor.
764        self._fh = fh
765        fn = None
766        try:
767            # XXX: Native calls to Windows filenos don't seem to work. 
768            #      Need to investigate why.
769            if not is_win32 and hasattr(fh, 'fileno'):
770                fn = fh.fileno()
771        except:
772            pass
773
774        if fn != None:
775            self.file = regfi.regfi_alloc(fn, REGFI_ENCODING_UTF8)
776            if not self.file:
777                # XXX: switch to non-generic exception
778                raise Exception("Could not open registry file.  Current log:\n"
779                                + getLogMessages())
780        else:
781            fh.seek(0)
782            self.raw_file = structures.REGFI_RAW_FILE()
783            self.raw_file.fh = fh
784            self.raw_file.seek = seek_cb_type(self.raw_file.cb_seek)
785            self.raw_file.read = read_cb_type(self.raw_file.cb_read)
786            self.file = regfi.regfi_alloc_cb(pointer(self.raw_file), REGFI_ENCODING_UTF8)
787            if not self.file:
788                # XXX: switch to non-generic exception
789                raise Exception("Could not open registry file.  Current log:\n"
790                                + getLogMessages())
791
792
793    def __getattr__(self, name):
794        if name == "root":
795            # XXX: This creates reference loops.  Need to cache better inside regfi
796            #if self._root == None:
797            #    self._root = Key(self, regfi.regfi_get_rootkey(self.file))
798            #return self._root
799            return Key(self, regfi.regfi_get_rootkey(self.file))
800
801        elif name == "modified":
802            return regfi.regfi_nt2unix_time(self._base.contents.mtime)
803
804        return getattr(self.file.contents, name)
805
806   
807    def __del__(self):
808        if self.file:
809            regfi.regfi_free(self.file)
810
811    def __iter__(self):
812        return HiveIterator(self)
813
814
815    ## Creates a @ref HiveIterator initialized at the specified path in
816    #  the hive.
817    #
818    # @param path A list of Key names which represent an absolute path within
819    #             the Hive
820    #
821    # @return A @ref HiveIterator which is positioned at the specified path.
822    #
823    # @exception Exception If the path could not be found/traversed
824    def subtree(self, path):
825        hi = HiveIterator(self)
826        hi.descend(path)
827        return hi
828
829
830## A special purpose iterator for registry hives
831#
832# Iterating over an object of this type causes all keys in a specific
833# hive subtree to be returned in a depth-first manner. These iterators
834# are typically created using the @ref Hive.subtree() function on a @ref Hive
835# object.
836#
837# HiveIterators can also be used to manually traverse up and down a
838# registry hive as they retain information about the current position in
839# the hive, along with which iteration state for subkeys and values for
840# every parent key.  See the @ref up and @ref down methods for more
841# information.
842class HiveIterator():
843    _hive = None
844    _iter = None
845    _iteration_root = None
846    _lock = None
847
848    def __init__(self, hive):
849        self._iter = regfi.regfi_iterator_new(hive.file)
850        if not self._iter:
851            raise Exception("Could not create iterator.  Current log:\n"
852                            + getLogMessages())
853        self._hive = hive
854        self._lock = threading.RLock()
855   
856    def __getattr__(self, name):
857        self._lock.acquire()
858        ret_val = getattr(self._iter.contents, name)
859        self._lock.release()
860        return ret_val
861
862    def __del__(self):
863        self._lock.acquire()
864        regfi.regfi_iterator_free(self._iter)
865        self._lock.release()
866
867    def __iter__(self):
868        self._lock.acquire()
869        self._iteration_root = None
870        self._lock.release()
871        return self
872
873    def __next__(self):
874        self._lock.acquire()
875        if self._iteration_root == None:
876            self._iteration_root = self.current_key().offset
877        elif not regfi.regfi_iterator_down(self._iter):
878            up_ret = regfi.regfi_iterator_up(self._iter)
879            while (up_ret and
880                   not regfi.regfi_iterator_next_subkey(self._iter)):
881                if self._iteration_root == self.current_key().offset:
882                    self._iteration_root = None
883                    self._lock.release()
884                    raise StopIteration('')
885                up_ret = regfi.regfi_iterator_up(self._iter)
886
887            if not up_ret:
888                self._iteration_root = None
889                self._lock.release()
890                raise StopIteration('')
891           
892            # XXX: Use non-generic exception
893            if not regfi.regfi_iterator_down(self._iter):
894                self._lock.release()
895                raise Exception('Error traversing iterator downward.'+
896                                ' Current log:\n'+ getLogMessages())
897
898        regfi.regfi_iterator_first_subkey(self._iter)
899        ret_val = self.current_key()
900        self._lock.release()
901
902        return ret_val
903
904
905    # For Python 2.x
906    next = __next__
907
908    # XXX: Should add sanity checks on some of these traversal functions
909    #      to throw exceptions if a traversal/retrieval *should* have worked
910    #      but failed for some reason.
911
912    ## Descends the iterator to a subkey
913    #
914    # Descends the iterator one level to the current subkey, or a subkey
915    # specified by name.
916    #
917    # @param subkey_name If specified, locates specified subkey by name
918    #                    (via find_subkey()) and descends to it.
919    #
920    # @return True if successful, False otherwise
921    def down(self, subkey_name=None):
922        ret_val = None
923        if subkey_name == None:
924            self._lock.acquire()
925            ret_val = regfi.regfi_iterator_down(self._iter)
926        else:
927            if name != None:
928                name = name.encode('utf-8')
929            self._lock.acquire()
930            ret_val = (regfi.regfi_iterator_find_subkey(self._iter, name) 
931                       and regfi.regfi_iterator_down(self._iter))
932       
933        self._lock.release()
934        return ret_val
935
936
937    ## Causes the iterator to ascend to the current Key's parent
938    #
939    # @return True if successful, False otherwise
940    #
941    # @note The state of current subkeys and values at this level in the tree
942    #       is lost as a side effect.  That is, if you go up() and then back
943    #       down() again, current_subkey() and current_value() will return
944    #       default selections.
945    def up(self):
946        self._lock.acquire()
947        ret_val = regfi.regfi_iterator_up(self._iter)
948        self._lock.release()
949        return ret_val
950
951
952    ## Selects first subkey of current key
953    #
954    # @return A Key instance for the first subkey. 
955    #         None on error or if the current key has no subkeys.
956    def first_subkey(self):
957        ret_val = None
958        self._lock.acquire()
959        if regfi.regfi_iterator_first_subkey(self._iter):
960            ret_val = self.current_subkey()
961        self._lock.release()
962        return ret_val
963
964
965    ## Selects first value of current Key
966    #
967    # @return A Value instance for the first value. 
968    #         None on error or if the current key has no values.
969    def first_value(self):
970        ret_val = None
971        self._lock.acquire()
972        if regfi.regfi_iterator_first_value(self._iter):
973            ret_val = self.current_value()
974        self._lock.release()
975        return ret_val
976
977
978    ## Selects the next subkey in the current Key's list
979    #
980    # @return A Key instance for the next subkey.
981    #         None if there are no remaining subkeys or an error occurred.
982    def next_subkey(self):
983        ret_val = None
984        self._lock.acquire()
985        if regfi.regfi_iterator_next_subkey(self._iter):
986            ret_val = self.current_subkey()
987        self._lock.release()
988        return ret_val
989
990
991    ## Selects the next value in the current Key's list
992   
993    # @return A Value instance for the next value.
994    #         None if there are no remaining values or an error occurred.
995    def next_value(self):
996        ret_val = None
997        self._lock.acquire()
998        if regfi.regfi_iterator_next_value(self._iter):
999            ret_val = self.current_value()
1000        self._lock.release()
1001        return ret_val
1002
1003
1004    ## Selects the first subkey which has the specified name
1005    #
1006    # @return A Key instance for the selected key.
1007    #         None if it could not be located or an error occurred.
1008    def find_subkey(self, name):
1009        if name != None:
1010            name = name.encode('utf-8')
1011        ret_val = None
1012        self._lock.acquire()
1013        if regfi.regfi_iterator_find_subkey(self._iter, name):
1014            ret_val = self.current_subkey()
1015        self._lock.release()
1016        return ret_val
1017
1018
1019    ## Selects the first value which has the specified name
1020    #
1021    # @return A Value instance for the selected value.
1022    #         None if it could not be located or an error occurred.
1023    def find_value(self, name):
1024        if name != None:
1025            name = name.encode('utf-8')
1026        ret_val = None
1027        self._lock.acquire()
1028        if regfi.regfi_iterator_find_value(self._iter, name):
1029            ret_val = self.current_value()
1030        self._lock.release()
1031        return ret_val
1032
1033    ## Retrieves the currently selected subkey
1034    #
1035    # @return A Key instance of the current subkey
1036    def current_subkey(self):
1037        self._lock.acquire()
1038        ret_val = Key(self._hive, regfi.regfi_iterator_cur_subkey(self._iter))
1039        self._lock.release()
1040        return ret_val
1041
1042    ## Retrieves the currently selected value
1043    #
1044    # @return A Value instance of the current value
1045    def current_value(self):
1046        self._lock.acquire()
1047        ret_val = Value(self._hive, regfi.regfi_iterator_cur_value(self._iter))
1048        self._lock.release()
1049        return ret_val
1050
1051    ## Retrieves the current key
1052    #
1053    # @return A Key instance of the current position of the iterator
1054    def current_key(self):
1055        self._lock.acquire()
1056        ret_val = Key(self._hive, regfi.regfi_iterator_cur_key(self._iter))
1057        self._lock.release()
1058        return ret_val
1059
1060    ## Traverse downward multiple levels
1061    #
1062    # This is more efficient than calling down() multiple times
1063    #
1064    # @param path A list of Key names which represent the path to descend
1065    #
1066    # @exception Exception If path could not be located
1067    def descend(self, path):
1068        cpath = _strlist2charss(path)
1069
1070        self._lock.acquire()
1071        result = regfi.regfi_iterator_descend(self._iter, cpath)
1072        self._lock.release()
1073        if not result:
1074            # XXX: Use non-generic exception
1075            raise Exception('Could not locate path.\n'+getLogMessages())
1076
1077    ## Obtains a list of the current key's ancestry
1078    #
1079    # @return A list of all parent keys starting with the root Key and ending
1080    #         with the current Key
1081    def ancestry(self):
1082        self._lock.acquire()
1083        result = regfi.regfi_iterator_ancestry(self._iter)
1084        self._lock.release()
1085
1086        ret_val = []
1087        i = 0
1088        k = result[i]
1089        while k:
1090            k = cast(regfi.regfi_reference_record(self._hive.file, k), POINTER(REGFI_NK))
1091            ret_val.append(Key(self._hive, k))
1092            i += 1
1093            k = result[i]
1094
1095        regfi.regfi_free_record(self._hive.file, result)
1096        return ret_val
1097
1098    ## Obtains the current path of the iterator
1099    #
1100    # @return A list of key names starting with the root up to and
1101    #         including the current key
1102    #
1103    def current_path(self):
1104        ancestry = self.ancestry()
1105        return [str(a.name) for a in ancestry]
1106
1107
1108# Freeing symbols defined for the sake of documentation
1109del Value.name,Value.name_raw,Value.offset,Value.data_size,Value.type,Value.flags
1110del Key.name,Key.name_raw,Key.offset,Key.modified,Key.flags
1111del Hive.root,Hive.modified,Hive.sequence1,Hive.sequence2,Hive.major_version,Hive.minor_version
1112del Security.ref_count,Security.offset,Security.descriptor
Note: See TracBrowser for help on using the repository browser.