from __future__ import generators

# This module is part of the spambayes project, which is Copyright 2002-3
# The Python Software Foundation and is covered by the Python Software
# Foundation license.

# Simple Python library for Outlook Express mailboxes handling
# Based on C++ work by Arne Schloh <oedbx@aroh.de>

__author__ = "Romain Guy"
__credits__ = "All the SpamBayes folk"

import binascii
import os
import struct
import msgs
import StringIO
import sys
from time import gmtime, strftime

try:
    import win32api
    import win32con
    from win32com.shell import shell, shellcon
except ImportError:
    # Not win32, or win32all not installed.
    # Some functions will not work, but some will.
    win32api = win32con = shell = shellcon = None

###########################################################################
## DBX FILE HEADER
###########################################################################
class dbxFileHeader:
    """
       Each Outlook Express DBX file has a file header.
       This header defines many properties, only a few of which interest us.

       The only properties which are required are defined by indexes. The
       indexes are static attributes of the class and their names begin with
       "fh". You can access their values through the method getEntry().
    """

    HEADER_SIZE    = 0x24bc              # total header size
    HEADER_ENTRIES = HEADER_SIZE >> 2    # total of entries in the header

    MAGIC_NUMBER   = 0xfe12adcfL         # specific to DBX files
    OFFLINE        = 0x26fe9d30L         # specific to offline.dbx
    FOLDERS        = 0x6f74fdc6L         # specific to folders.dbx
    POP3UIDL       = 0x6f74fdc7L         # specific to pop3uidl.dbx

    # various entries indexes
    FH_FILE_INFO_LENGTH       = 0x07     # file info length
    FH_FIRST_FOLDER_LIST_NODE = 0x1b     # pointer to the first folder list node
    FH_LAST_FOLDER_LIST_NODE  = 0x1c     # pointer to the last folder list node
    FH_MESSAGE_CONDITIONS_PTR = 0x22     # pointer to the message conditions object
    FH_FOLDER_CONDITIONS_PTR  = 0x23     # pointer to the folder conditions object
    FH_ENTRIES                = 0x31     # entries in tree
    FH_TREE_ROOT_NODE_PTR     = 0x39     # pointer to the root node of a tree

    FILE_HEADER_ENTRIES = \
    [ ( 0x07,  "file info length"                                ),
      ( 0x09,  "pointer to the last variable segment"            ),
      ( 0x0a,  "length of a variable segment"                    ),
      ( 0x0b,  "used space of the last variable segment"         ),
      ( 0x0c,  "pointer to the last tree segment"                ),
      ( 0x0d,  "length of a tree segment"                        ),
      ( 0x0e,  "used space of the last tree segment"             ),
      ( 0x0f,  "pointer to the last message segment"             ),
      ( 0x10,  "length of a message segment"                     ),
      ( 0x11,  "used space of the last message segment"          ),
      ( 0x12,  "root pointer to the deleted message list"        ),
      ( 0x13,  "root pointer to the deleted tree list"           ),
      ( 0x15,  "used space in the middle sector of the file"     ),
      ( 0x16,  "reusable space in the middle sector of the file" ),
      ( 0x17,  "index of the last entry in the tree"             ),
      ( 0x1b,  "pointer to the first folder list node"           ),
      ( 0x1c,  "pointer to the last folder list node"            ),
      ( 0x1f,  "used space of the file"                          ),
      ( 0x22,  "pointer to the message conditions object"        ),
      ( 0x23,  "pointer to the folder conditions object"         ),
      ( 0x31,  "entries in the tree"                             ),
      ( 0x32,  "entries in the 2.nd tree"                        ),
      ( 0x33,  "entries in the 3.rd tree"                        ),
      ( 0x39,  "pointer to the root node of the tree"            ),
      ( 0x3a,  "pointer to the root node of the 2.nd tree"       ),
      ( 0x3b,  "pointer to the root node of the 3.rd tree"       ),
      ( 0x9f,  "used space for indexed info objects"             ),
      ( 0xa0,  "used space for conditions objects"               ),
      ( 0xa2,  "used space for folder list objects"              ),
      ( 0xa3,  "used space for tree objects"                     ),
      ( 0xa4,  "used space for message objects"                  )]

    def __init__(self, dbxStream):
        """Initialize the DBX header by reading it directly from the passed
        stream."""
        dbxStream.seek(0)
        self.dbxBuffer = dbxStream.read(dbxFileHeader.HEADER_SIZE)

    def isMessages(self):
        """Return true iff the DBX is a messages DBX."""
        return not (self.isFolders() or self.isPOP3UIDL() or self.isOffline())

    def isFolders(self):
        """Return true if the DBX is the folders DBX."""
        return self.getEntry(1) == dbxFileHeader.FOLDERS

    def isPOP3UIDL(self):
        """Return true if the DBX is the POP3UIDL DBX."""
        return self.getEntry(1) == dbxFileHeader.POP3UIDL

    def isOffline(self):
        """Return true if the DBX is the offline DBX."""
        return self.getEntry(1) == dbxFileHeader.OFFLINE

    def isValid(self):
        """Return true if the DBX is a valid DBX file."""
        return self.getEntry(0) == dbxFileHeader.MAGIC_NUMBER

    def getHeaderBuffer(self):
        """Return the bytes buffer containing the whole header."""
        return self.dbxBuffer

    def getEntry(self, dbxEntry):
        """Return the n-th entry as a long integer."""
        return struct.unpack("L",
                             self.dbxBuffer[dbxEntry * 4:(dbxEntry * 4) + 4])[0]

    def getEntryAsHexStr(self, dbxEntry):
        """Return the n-th entry as an hexadecimal string.
        (Little endian encoding!)"""
        return '0x' + \
               binascii.hexlify(self.dbxBuffer[dbxEntry * 4:(dbxEntry * 4) + 4])

###########################################################################
## DBX FILE INFO
###########################################################################

class dbxFileInfo:
    """
      Following the DBX header there is DBX info. This part gives the name of
      the folder described by the current DBX.
    """

    MESSAGE_FILE_INFO = 0x618

    def __init__(self, dbxStream, dbxLength):
        """Reads the DBX info part from a DBX stream."""
        dbxStream.seek(dbxFileHeader.HEADER_SIZE)
        self.dbxLength = dbxLength
        self.dbxBuffer = dbxStream.read(dbxLength)

    def isFoldersInfo(self):
        """Return true if the info belongs to folders.dbx."""
        return self.dbxLength != dbxFileInfo.MESSAGE_FILE_INFO

    def getFolderName(self):
        """Returns the folder name."""
        if not self.isFoldersInfo():
            name = [c for c in self.dbxBuffer[0x105:0x210] if ord(c) != 0]
            return "".join(name)
        else:
            return None

    def getCreationTime(self):
        """Not implemented yet."""
        if self.isFoldersInfo():
            return "Not implemented yet"
        else:
            return None

###########################################################################
## DBX TREE
###########################################################################

class dbxTree:
    """Stands for the tree which stores the messages in a given folder."""

    TREE_NODE_SIZE = 0x27c          # size of a tree node

    def __init__(self, dbxStream, dbxAddress, dbxValues):
        """Reads the addresses of the stored messages."""
        self.dbxValues = [i for i in range(dbxValues)]
        # XXX : silly fix !
        if dbxAddress > 0:
            self.__readValues(dbxStream, 0, dbxAddress, 0, dbxValues)

    def __readValues(self, dbxStream, dbxParent, dbxAddress, dbxPosition, dbxValues):
        dbxStream.seek(dbxAddress)
        dbxBuffer = dbxStream.read(dbxTree.TREE_NODE_SIZE)

        count = 0
        entries = ((self.getEntry(dbxBuffer, 4) >> 8) & 0xff)

        if self.getEntry(dbxBuffer, 2) != 0:
            self.__readValues(dbxStream, dbxAddress, self.getEntry(dbxBuffer, 2),
                                         dbxPosition, self.getEntry(dbxBuffer, 5))
            count += self.getEntry(dbxBuffer, 5)

        for i in range(entries):
            pos = 6 + i * 3

            if self.getEntry(dbxBuffer, pos) != 0:
                count += 1
                value = dbxPosition + count
                self.dbxValues[value - 1] = self.getEntry(dbxBuffer, pos)

            if self.getEntry(dbxBuffer, pos + 1) != 0:
                self.__readValues(dbxStream, dbxAddress, self.getEntry(dbxBuffer, pos + 1),
                                             dbxPosition + count, self.getEntry(dbxBuffer, pos + 2))
                count += self.getEntry(dbxBuffer, pos + 2)

    def getEntry(self, dbxBuffer, dbxEntry):
        """Return the n-th entry as a long integer."""
        return struct.unpack("L", dbxBuffer[dbxEntry * 4:(dbxEntry * 4) + 4])[0]

    def getValue(self, dbxIndex):
        """Return the address of the n-th message."""
        return self.dbxValues[dbxIndex]

###########################################################################
## DBX INDEXED INFO
###########################################################################

class dbxIndexedInfo:
    """
      Messages and folders mailboxes contain the "message info" and "folders
      info" entities.
      These entities are indexed info sequences. This is their base class.
    """

    MAX_INDEX = 0x20        # max index
    DT_NONE   = 0           # data type none

    def __init__(self, dbxStream, dbxAddress):
        """Reads the indexed infos from the passed stream."""
        self.dbxBodyLength   = 0L
        self.dbxObjectLength = 0L
        self.dbxEntries      = 0L
        self.dbxCounter      = 0L
        self.dbxBuffer       = []
        self.dbxIndexes      = 0L
        self.dbxBegin        = [0L for i in range(dbxIndexedInfo.MAX_INDEX)]
        self.dbxLength       = [i  for i in self.dbxBegin]
        self.dbxAddress       = dbxAddress
        self.__readIndexedInfo(dbxStream)

    def __readIndexedInfo(self, dbxStream):
        dbxStream.seek(self.dbxAddress)
        temp = dbxStream.read(12)

        self.dbxBodyLength   =  self.__getEntry(temp, 1)
        self.dbxObjectLength =  self.__getEntry(temp, 2) & 0xffff
        self.dbxEntries      = (self.__getEntry(temp, 2) >> 16) & 0xff
        self.dbxCounter      = (self.__getEntry(temp, 1) >> 24) & 0xff
        self.dbxBuffer       =  dbxStream.read(self.dbxBodyLength) # bytes array

        isIndirect           = bool(0)                             # boolean
        lastIndirect         = 0
        data                 = self.dbxEntries << 2                # index within dbxBuffer

        for i in range(self.dbxEntries):
            value      = self.__getEntry(self.dbxBuffer, i)
            isDirect   = value & 0x80
            index      = value & 0x7f
            value    >>= 8
            if isDirect:
                self.__setIndex(index, (i << 2) + 1, 3)
            else:
                self.__setIndex(index, data + value)
                if isIndirect:
                    self.__setEnd(lastIndirect, data + value)
                    isIndirect   = bool(1)
                    lastIndirect = index
            self.dbxIndexes |= 1 << index

        if isIndirect:
            self.__setEnd(lastIndirect, self.dbxBodyLength)

    def __setIndex(self, dbxIndex, dbxBegin, dbxLength = 0):
        if dbxIndex < dbxIndexedInfo.MAX_INDEX:
            self.dbxBegin[dbxIndex] = dbxBegin
            self.dbxLength[dbxIndex] = dbxLength

    def __setEnd(self, dbxIndex, dbxEnd):
        if dbxIndex < dbxIndexedInfo.MAX_INDEX:
            self.dbxLength[dbxIndex] = dbxEnd - self.dbxBegin[dbxIndex]

    def __getEntry(self, dbxBuffer, dbxEntry):
        return struct.unpack("L", dbxBuffer[dbxEntry * 4:(dbxEntry * 4) + 4])[0]

    def getIndexText(self, dbxIndex):
        """Returns the description of the given indexed field."""
        return ""

    def getIndexDataType(self, dbxIndex):
        """Returns the data type of the given index."""
        return DT_NONE

    def getValue(self, dbxIndex):
        """Returns a tuple : (index in buffer of the info, length of the info)."""
        return (self.dbxBegin[dbxIndex], self.dbxLength[dbxIndex])

    def getValueAsLong(self, dbxIndex):
        """Returns the indexed info as a long value."""
        data, length = self.getValue(dbxIndex)
        value  = 0
        if data:
            value = struct.unpack("L", self.dbxBuffer[data:data + 4])[0]
            if length < 4:
                value &= (1 << (length << 3)) - 1
        return value

    def getString(self, dbxIndex):
        """Returns the indexed info as a string value."""
        index = self.dbxBegin[dbxIndex]
        end = index
        for c in self.dbxBuffer[index:]:
            if ord(c) == 0: break
            end += 1
        return self.dbxBuffer[index:end]

    def getAddress(self):
        return self.dbxAddress

    def getBodyLength(self):
        return self.dbxBodyLength

    def getEntries(self):
        return self.dbxEntries

    def getCounter(self):
        return self.dbxCounter

    def getIndexes(self):
        return self.dbxIndexes

    def isIndexed(self, dbxIndex):
        return self.dbxIndexes & (1 << dbxIndex)

###########################################################################
## DBX MESSAGE INFO
###########################################################################

class dbxMessageInfo(dbxIndexedInfo):
    """
      The message info structure inherits from the index info one. It just
      defines extra constants which allow to access pertinent info.
    """

    MI_INDEX           = 0x0                   # index of the message
    MI_FLAGS           = 0x1                   # the message flags
    MI_MESSAGE_ADDRESS = 0x4                   # the address of the message
    MI_SUBJECT         = 0x8                   # the subject of the message

    # label of each indexed info
    INDEX_LABEL = \
    [ "message index"                , "flags"                          ,
      "time message created/send"    , "body lines"                     ,
      "message address"              , "original subject"               ,
      "time message saved"           , "message id"                     ,
      "subject"                      , "sender eMail address and name"  ,
      "answered to message id"       , "server/newsgroup/message number",
      "server"                       , "sender name"                    ,
      "sender eMail address"         , "id 0f"                          ,
      "message priority"             , "message text length"            ,
      "time message created/received", "receiver name"                  ,
      "receiver eMail address"       , "id 15"                          ,
      "id 16"                        , "id 17"                          ,
      "id 18"                        , "id 19"                          ,
      "OE account name"              , "OE account registry key"        ,
      "message text structure"       , "id 1d"                          ,
      "id 1e"                        , "id 1f"                           ]

    DT_NONE      = 0                    # index is none
    DT_INT4      = 1                    # index is a long integer (32 bits)
    DT_STRING    = 2                    # index is a string
    DT_DATE_TIME = 3                    # index is date/time
    DT_DATA      = 4                    # index is data

    # the data type of each index
    INDEX_DATA_TYPE = \
    [ DT_INT4  , DT_INT4  , DT_DATE_TIME, DT_INT4  , DT_INT4  , DT_STRING, DT_DATE_TIME, DT_STRING,
      DT_STRING, DT_STRING, DT_STRING   , DT_STRING, DT_STRING, DT_STRING, DT_STRING   , DT_NONE  ,
      DT_INT4  , DT_INT4  , DT_DATE_TIME, DT_STRING, DT_STRING, DT_NONE  , DT_INT4     , DT_NONE  ,
      DT_INT4  , DT_INT4  , DT_STRING   , DT_STRING, DT_DATA  , DT_NONE  , DT_NONE     , DT_NONE   ]

    def getIndexText(self, dbxIndex):
        return dbxMessageInfo.INDEX_LABEL[dbxIndex]

    def getIndexDataType(self, dbxIndex):
        return dbxMessageInfo.INDEX_DATA_TYPE[dbxIndex]

###########################################################################
## DBX MESSAGE
###########################################################################

class dbxMessage:
    def __init__(self, dbxStream, dbxAddress):
        self.dbxAddress = dbxAddress
        self.dbxText   = ""
        self.dbxLength = 0L
        self.__readMessageText(dbxStream)

    def __getEntry(self, dbxBuffer, dbxEntry):
        if len(dbxBuffer) < (dbxEntry * 4) + 4:
            return None
        return struct.unpack("L", dbxBuffer[dbxEntry * 4:(dbxEntry * 4) + 4])[0]

    def __readMessageText(self, dbxStream):
        address = self.dbxAddress
        header = ""

        while (address):
            dbxStream.seek(address)
            header = dbxStream.read(16)

            self.dbxLength += self.__getEntry(header, 2)
            address          = self.__getEntry(header, 3)

        pos = ""
        address = self.dbxAddress

        while (address):
            dbxStream.seek(address)
            header  = dbxStream.read(16)
            pos    += dbxStream.read(self.__getEntry(header, 2))
            address  = self.__getEntry(header, 3)

        self.dbxText = pos

    def getText(self):
        return self.dbxText

# This started its SpamBayes life as a private method of the UserInterface
# class, but is really a general purpose (Outlook Express) function.
def convertToMbox(content):
    """Check if the given buffer is in a non-mbox format, and convert it
    into mbox format if so.  If it's already an mbox, return it unchanged.
    """

    dbxStream = StringIO.StringIO(content)
    header = dbxFileHeader(dbxStream)

    if header.isValid() and header.isMessages():
        file_info_len = dbxFileHeader.FH_FILE_INFO_LENGTH
        fh_entries = dbxFileHeader.FH_ENTRIES
        fh_ptr = dbxFileHeader.FH_TREE_ROOT_NODE_PTR

        info = dbxFileInfo(dbxStream, header.getEntry(file_info_len))
        entries = header.getEntry(fh_entries)
        address = header.getEntry(fh_ptr)

        if address and entries:
            tree = dbxTree(dbxStream, address, entries)
            dbxBuffer = ""

            for i in range(entries):
                address = tree.getValue(i)
                messageInfo = dbxMessageInfo(dbxStream, address)

                if messageInfo.isIndexed(dbxMessageInfo.MI_MESSAGE_ADDRESS):
                    address = dbxMessageInfo.MI_MESSAGE_ADDRESS
                    messageAddress = messageInfo.getValueAsLong(address)
                    message = dbxMessage(dbxStream, messageAddress)

                    # This fakes up a from header to conform to mbox
                    # standards.  It would be better to extract this
                    # data from the message itself, as this will
                    # result in incorrect tokens.
                    dbxBuffer += "From spambayes@spambayes.org %s\n%s" \
                                 % (strftime("%a %b %d %H:%M:%S MET %Y",
                                             gmtime()), message.getText())
            content = dbxBuffer
    dbxStream.close()
    return content

def OEIdentityKeys():
    """Return the OE identity keys.

    Tested with Outlook Express 6.0 with Windows XP."""
    if sys.platform != "win32":
        # AFAIK, there is only a Win32 OE, and a Mac OE.
        # The Mac OE should be easy enough, but I don't know
        # where the dbx files are stored (I presume they are in the
        # same format).
        raise NotImplementedError
    if win32api is None:
        # Delayed import error from top.
        raise ImportError("win32all not installed")

    reg = win32api.RegOpenKeyEx(win32con.HKEY_USERS, "")
    user_index = 0
    while True:
        # Loop through all the users
        try:
            user_name = "%s\\Identities" % \
                        (win32api.RegEnumKey(reg, user_index),)
        except win32api.error:
            break
        user_index += 1
        try:
            user_key = win32api.RegOpenKeyEx(win32con.HKEY_USERS, user_name)
        except win32api.error:
            # Not this one
            continue

        identity_index = 0
        while True:
            # Loop through all the identities
            try:
                identity_name = win32api.RegEnumKey(user_key,
                                                    identity_index)
            except win32api.error:
                break
            identity_index += 1

            subkey_name = "%s\\%s\\%s" % (user_name, identity_name,
                                          "Software\\Microsoft\\Outlook " \
                                          "Express\\5.0")
            try:
                subkey = win32api.RegOpenKeyEx(win32con.HKEY_USERS,
                                               subkey_name, 0,
                                               win32con.KEY_READ)
            except win32api.error:
                # Not this user
                continue
            yield subkey

def OEStoreRoot():
    """Return the path to the Outlook Express Store Root.

    Tested with Outlook Express 6.0 with Windows XP."""
    # Run through the identity keys, using the first that
    # works.
    raw = ""
    for identity in OEIdentityKeys():
        try:
            raw = win32api.RegQueryValueEx(identity, "Store Root")
        except win32api.error:
            pass
        else:
            break
    # I can't find a shellcon to that is the same as %UserProfile%,
    # so extract it from CSIDL_LOCAL_APPDATA
    UserDirectory = shell.SHGetFolderPath \
                    (0, shellcon.CSIDL_LOCAL_APPDATA, 0, 0)
    parts = UserDirectory.split(os.sep)
    UserProfile = os.sep.join(parts[:-2])
    raw = raw[0].replace("%UserProfile%", UserProfile)
    return raw

def OEAccountKeys(permission = None):
    """Return registry keys for each of the OE mail accounts, along
    with information about what type of mail account it is."""
    if permission is None:
        # Can't do this in the parameter, because then it requires
        # win32con to be available for the module to be imported.
        permission = win32con.KEY_READ | win32con.KEY_SET_VALUE

    possible_root_keys = []

    # This appears to be the place for OE6 and WinXP
    # (So I'm guessing also for NT4)
    if sys.getwindowsversion()[0] >= 4:
        possible_root_keys = ["Software\\Microsoft\\" \
                             "Internet Account Manager\\Accounts"]
    else:
        # This appears to be the place for OE6 and Win98
        # (So I'm guessing also for Win95)
        possible_root_keys = oe_mailbox.OEIdentityKeys()

    for key in possible_root_keys:
        reg = win32api.RegOpenKeyEx(win32con.HKEY_CURRENT_USER, key)
        account_index = 0
        while True:
            # Loop through all the accounts
            account = {}
            try:
                subkey_name = "%s\\%s" % \
                              (key, win32api.RegEnumKey(reg, account_index))
            except win32api.error:
                break
            account_index += 1
            index = 0
            subkey = win32api.RegOpenKeyEx(win32con.HKEY_CURRENT_USER,
                                           subkey_name, 0, permission)
            while True:
                # Loop through all the keys so that we can determine
                # what type of account this is.
                try:
                    name, value, typ = win32api.RegEnumValue(subkey, index)
                except win32api.error:
                    break
                account[name] = (value, typ)
                index += 1

            # Yield, as appropriate.
            if account.has_key("POP3 Server"):
                yield("POP3", subkey, account)
            elif account.has_key("IMAP Server"):
                yield("IMAP4", subkey, account)

def OEIsInstalled():
    """Return True if Outlook Express appears to be installed,
    and in use (I think if sys.platform == "win32" would say if
    it was installed at all)."""
    # Our heuristic is that there is at least one mail account setup.
    if len(list(OEAccountKeys)) > 0:
        return True
    return False

## For use by the test tools.
class OEMsg(msgs.Msg):
    def __init__(self, guts, id):
        self.tag = id
        self.guts = guts

# The iterator yields a stream of Msg objects, taken from a list of
# dbx files.
class OEMsgStream(msgs.MsgStream):
    def __init__(self, tag, dbxes, keep=None):
        msgs.MsgStream.__init__(self, tag, dbxes, keep)

    def produce(self):
        if self.keep is None:
            for dbx in self.directories:
                folder = convertToMbox(file(dbx))
                all = folder.split("\nFrom ") # XXX Is this right?
                count = 0
                for msg in all:
                    id = "%s::%s" % (dbx, count)
                    count += 1
                    yield OEMsg(msg, id)
            return
        # We only want part of the msgs.  Shuffle each directory list, but
        # in such a way that we'll get the same result each time this is
        # called on the same directory list.
        for directory in self.directories:
            folder = convertToMbox(file(dbx))
            all = folder.split("\nFrom ") # XXX Is this right?
            random.seed(hash(max(all)) ^ SEED) # reproducible across calls
            random.shuffle(all)
            del all[self.keep:]
            all.sort()  # for consistency with MsgStream
            count = 0
            for msg in all:
                id = "%s::%s" % (dbx, count)
                count += 1
                yield OEMsg(msg, id)

class OEHamStream(msgs.HamStream):
    def __init__(self, tag, dbxes, train=0):
        msgs.HamStream.__init__(self, tag, dbxes, train)

class OESpamStream(msgs.SpamStream):
    def __init__(self, tag, dbxes, train=0):
        msgs.SpamStream.__init__(self, tag, dbxes, train)


###########################################################################
## TEST DRIVER
###########################################################################

def test():
    import sys
    import getopt

    try:
        opts, args = getopt.getopt(sys.argv[1:], 'hp')
    except getopt.error, msg:
        print >>sys.stderr, str(msg) + '\n\n' + __doc__
        sys.exit()

    print_message = False
    for opt, arg in opts:
        if opt == '-h':
            print >>sys.stderr, __doc__
            sys.exit()
        elif opt == '-p':
            print_message = True

    if args:
        MAILBOX_DIR = args[0]
    else:
        MAILBOX_DIR = OEStoreRoot()

    files = [os.path.join(MAILBOX_DIR, file) for file in \
             os.listdir(MAILBOX_DIR) if os.path.splitext(file)[1] == '.dbx']

    for file in files:
        try:
            print
            print file

            dbx = open(file, "rb", 0)
            header = dbxFileHeader(dbx)

            print "IS VALID DBX  :", header.isValid()

            if header.isMessages():
                info = dbxFileInfo(dbx, header.getEntry(dbxFileHeader.FH_FILE_INFO_LENGTH))
                print "MAILBOX NAME  :", info.getFolderName()
                print "CREATION TIME :", info.getCreationTime()

                entries = header.getEntry(dbxFileHeader.FH_ENTRIES)
                address  = header.getEntry(dbxFileHeader.FH_TREE_ROOT_NODE_PTR)

                if address and entries:
                    tree = dbxTree(dbx, address, entries)

                for i in range(entries):
                    address = tree.getValue(i)
                    messageInfo = dbxMessageInfo(dbx, address)

                    if messageInfo.isIndexed(dbxMessageInfo.MI_MESSAGE_ADDRESS):
                        messageAddress = messageInfo.getValueAsLong(dbxMessageInfo.MI_MESSAGE_ADDRESS)
                        message        = dbxMessage(dbx, messageAddress)

                        if print_message:
                            print
                            print "Message :", messageInfo.getString(dbxMessageInfo.MI_SUBJECT)
                            print "=" * (len(messageInfo.getString(dbxMessageInfo.MI_SUBJECT)) + 9)
                            print
                            print message.getText()

        except Exception, (strerror):
            print strerror

        dbx.close()

if __name__ == '__main__':
    test()


syntax highlighted by Code2HTML, v. 0.9.1