python-pillow/Pillow
1
1
Fork 0
mirror of https://github.com/python-pillow/Pillow.git synced 2025-01-28 02:04:36 +03:00
Code Issues Packages Projects Releases Wiki Activity

Port more stuff to Python 3, dropping some pre-2.6 support

Browse Source
This commit is contained in:
Martin Panter 2014-01-17 11:58:34 +00:00
parent cd61720845
commit 6977b592e3
1 changed files with 129 additions and 155 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

248
PIL/OleFileIO.py
View File

@ -171,7 +171,7 @@ __version__ = '0.26'
# - create a simple OLE explorer with wxPython # - create a simple OLE explorer with wxPython
# FUTURE EVOLUTIONS to add write support: # FUTURE EVOLUTIONS to add write support:
# 1) add ability to write a stream back on disk from StringIO (same size, no # 1) add ability to write a stream back on disk from BytesIO (same size, no
# change in FAT/MiniFAT). # change in FAT/MiniFAT).
# 2) rename a stream/storage if it doesn't change the RB tree # 2) rename a stream/storage if it doesn't change the RB tree
# 3) use rbtree module to update the red-black tree + any rename # 3) use rbtree module to update the red-black tree + any rename
@ -223,8 +223,7 @@ __version__ = '0.26'
import io import io
import sys import sys
from PIL import _binary from PIL import _binary
from PIL._util import isPath import struct, array, os.path, datetime
import struct, array, os.path, sys, datetime
#[PL] Define explicitly the public API to avoid private objects in pydoc: #[PL] Define explicitly the public API to avoid private objects in pydoc:
__all__ = ['OleFileIO', 'isOleFile'] __all__ = ['OleFileIO', 'isOleFile']
@ -240,19 +239,13 @@ elif array.array('I').itemsize == 4:
# on 64 bits platforms, integers in an array are 32 bits: # on 64 bits platforms, integers in an array are 32 bits:
UINT32 = 'I' UINT32 = 'I'
else: else:
raise ValueError, 'Need to fix a bug with 32 bit arrays, please contact author...' raise ValueError('Need to fix a bug with 32 bit arrays, please contact author...')
#[PL] These workarounds were inspired from the Path module #[PL] These workarounds were inspired from the Path module
# (see http://www.jorendorff.com/articles/python/path/) # (see http://www.jorendorff.com/articles/python/path/)
#TODO: test with old Python versions #TODO: test with old Python versions
# Pre-2.3 workaround for booleans
try:
True, False
except NameError:
True, False = 1, 0
# Pre-2.3 workaround for basestring. # Pre-2.3 workaround for basestring.
try: try:
basestring basestring
@ -271,7 +264,7 @@ KEEP_UNICODE_NAMES = False
# command line to change it. # command line to change it.
DEBUG_MODE = False DEBUG_MODE = False
def debug_print(msg): def debug_print(msg):
print msg print(msg)
def debug_pass(msg): def debug_pass(msg):
pass pass
debug = debug_pass debug = debug_pass
@ -292,15 +285,15 @@ def set_debug_mode(debug_mode):
MAGIC = b'\320\317\021\340\241\261\032\341' MAGIC = b'\320\317\021\340\241\261\032\341'
#[PL]: added constants for Sector IDs (from AAF specifications) #[PL]: added constants for Sector IDs (from AAF specifications)
MAXREGSECT = 0xFFFFFFFAL; # maximum SECT MAXREGSECT = 0xFFFFFFFA; # maximum SECT
DIFSECT = 0xFFFFFFFCL; # (-4) denotes a DIFAT sector in a FAT DIFSECT = 0xFFFFFFFC; # (-4) denotes a DIFAT sector in a FAT
FATSECT = 0xFFFFFFFDL; # (-3) denotes a FAT sector in a FAT FATSECT = 0xFFFFFFFD; # (-3) denotes a FAT sector in a FAT
ENDOFCHAIN = 0xFFFFFFFEL; # (-2) end of a virtual stream chain ENDOFCHAIN = 0xFFFFFFFE; # (-2) end of a virtual stream chain
FREESECT = 0xFFFFFFFFL; # (-1) unallocated sector FREESECT = 0xFFFFFFFF; # (-1) unallocated sector
#[PL]: added constants for Directory Entry IDs (from AAF specifications) #[PL]: added constants for Directory Entry IDs (from AAF specifications)
MAXREGSID = 0xFFFFFFFAL; # maximum directory entry ID MAXREGSID = 0xFFFFFFFA; # maximum directory entry ID
NOSTREAM = 0xFFFFFFFFL; # (-1) unallocated directory entry NOSTREAM = 0xFFFFFFFF; # (-1) unallocated directory entry
#[PL] object types in storage (from AAF specifications) #[PL] object types in storage (from AAF specifications)
STGTY_EMPTY = 0 # empty directory entry (according to OpenOffice.org doc) STGTY_EMPTY = 0 # empty directory entry (according to OpenOffice.org doc)
@ -348,7 +341,7 @@ DEFECT_FATAL = 40 # an error which cannot be ignored, parsing is
# impossible # impossible
#[PL] add useful constants to __all__: #[PL] add useful constants to __all__:
for key in vars().keys(): for key in list(vars().keys()):
if key.startswith('STGTY_') or key.startswith('DEFECT_'): if key.startswith('STGTY_') or key.startswith('DEFECT_'):
__all__.append(key) __all__.append(key)
@ -380,7 +373,7 @@ def _clsid(clsid):
clsid: string of length 16. clsid: string of length 16.
""" """
assert len(clsid) == 16 assert len(clsid) == 16
if clsid == "\0" * len(clsid): if clsid == bytearray(16):
return "" return ""
return (("%08X-%04X-%04X-%02X%02X-" + "%02X" * 6) % return (("%08X-%04X-%04X-%02X%02X-" + "%02X" * 6) %
((i32(clsid, 0), i16(clsid, 4), i16(clsid, 6)) + ((i32(clsid, 0), i16(clsid, 4), i16(clsid, 6)) +
@ -388,19 +381,15 @@ def _clsid(clsid):
# UNICODE support for Old Python versions: # UNICODE support:
# (necessary to handle storages/streams names which use Unicode) # (necessary to handle storages/streams names which use Unicode)
try:
# is Unicode supported ?
unicode
def _unicode(s, errors='replace'): def _unicode(s, errors='replace'):
""" """
Map unicode string to Latin 1. (Python with Unicode support) Map unicode string to Latin 1. (Python with Unicode support)
s: UTF-16LE unicode string to convert to Latin-1 s: UTF-16LE unicode string to convert to Latin-1
errors: 'replace', 'ignore' or 'strict'. See Python doc for unicode() errors: 'replace', 'ignore' or 'strict'.
""" """
#TODO: test if it OleFileIO works with Unicode strings, instead of #TODO: test if it OleFileIO works with Unicode strings, instead of
# converting to Latin-1. # converting to Latin-1.
@ -415,21 +404,7 @@ try:
return u.encode('latin_1', errors) return u.encode('latin_1', errors)
except: except:
# there was an error during Unicode to Latin-1 conversion: # there was an error during Unicode to Latin-1 conversion:
raise IOError, 'incorrect Unicode name' raise IOError('incorrect Unicode name')
except NameError:
def _unicode(s, errors='replace'):
"""
Map unicode string to Latin 1. (Python without native Unicode support)
s: UTF-16LE unicode string to convert to Latin-1
errors: 'replace', 'ignore' or 'strict'. (ignored in this version)
"""
# If the unicode function does not exist, we assume this is an old
# Python version without Unicode support.
# Null bytes are simply removed (this only works with usual Latin-1
# strings which do not contain unicode characters>256):
return filter(ord, s)
def filetime2datetime(filetime): def filetime2datetime(filetime):
@ -439,8 +414,8 @@ def filetime2datetime(filetime):
# TODO: manage exception when microseconds is too large # TODO: manage exception when microseconds is too large
# inspired from http://code.activestate.com/recipes/511425-filetime-to-datetime/ # inspired from http://code.activestate.com/recipes/511425-filetime-to-datetime/
_FILETIME_null_date = datetime.datetime(1601, 1, 1, 0, 0, 0) _FILETIME_null_date = datetime.datetime(1601, 1, 1, 0, 0, 0)
#debug('timedelta days=%d' % (filetime/(10*1000000*3600*24))) #debug('timedelta days=%d' % (filetime//(10*1000000*3600*24)))
return _FILETIME_null_date + datetime.timedelta(microseconds=filetime/10) return _FILETIME_null_date + datetime.timedelta(microseconds=filetime//10)
@ -588,14 +563,14 @@ class OleMetadata:
""" """
Dump all metadata, for debugging purposes. Dump all metadata, for debugging purposes.
""" """
print 'Properties from SummaryInformation stream:' print('Properties from SummaryInformation stream:')
for prop in self.SUMMARY_ATTRIBS: for prop in self.SUMMARY_ATTRIBS:
value = getattr(self, prop) value = getattr(self, prop)
print '- %s: %s' % (prop, repr(value)) print('- %s: %s' % (prop, repr(value)))
print 'Properties from DocumentSummaryInformation stream:' print('Properties from DocumentSummaryInformation stream:')
for prop in self.DOCSUM_ATTRIBS: for prop in self.DOCSUM_ATTRIBS:
value = getattr(self, prop) value = getattr(self, prop)
print '- %s: %s' % (prop, repr(value)) print('- %s: %s' % (prop, repr(value)))
#--- _OleStream --------------------------------------------------------------- #--- _OleStream ---------------------------------------------------------------
@ -605,7 +580,7 @@ class _OleStream(io.BytesIO):
OLE2 Stream OLE2 Stream
Returns a read-only file object which can be used to read Returns a read-only file object which can be used to read
the contents of a OLE stream (instance of the StringIO class). the contents of a OLE stream (instance of the BytesIO class).
To open a stream, use the openstream method in the OleFile class. To open a stream, use the openstream method in the OleFile class.
This function can be used with either ordinary streams, This function can be used with either ordinary streams,
@ -631,7 +606,7 @@ class _OleStream(io.BytesIO):
sectorsize: size of one sector sectorsize: size of one sector
fat : array/list of sector indexes (FAT or MiniFAT) fat : array/list of sector indexes (FAT or MiniFAT)
filesize : size of OLE file (for debugging) filesize : size of OLE file (for debugging)
return : a StringIO instance containing the OLE stream return : a BytesIO instance containing the OLE stream
""" """
debug('_OleStream.__init__:') debug('_OleStream.__init__:')
debug(' sect=%d (%X), size=%d, offset=%d, sectorsize=%d, len(fat)=%d, fp=%s' debug(' sect=%d (%X), size=%d, offset=%d, sectorsize=%d, len(fat)=%d, fp=%s'
@ -647,12 +622,12 @@ class _OleStream(io.BytesIO):
# and we keep a record that size was unknown: # and we keep a record that size was unknown:
unknown_size = True unknown_size = True
debug(' stream with UNKNOWN SIZE') debug(' stream with UNKNOWN SIZE')
nb_sectors = (size + (sectorsize-1)) / sectorsize nb_sectors = (size + (sectorsize-1)) // sectorsize
debug('nb_sectors = %d' % nb_sectors) debug('nb_sectors = %d' % nb_sectors)
# This number should (at least) be less than the total number of # This number should (at least) be less than the total number of
# sectors in the given FAT: # sectors in the given FAT:
if nb_sectors > len(fat): if nb_sectors > len(fat):
raise IOError, 'malformed OLE document, stream too large' raise IOError('malformed OLE document, stream too large')
# optimization(?): data is first a list of strings, and join() is called # optimization(?): data is first a list of strings, and join() is called
# at the end to concatenate all in one string. # at the end to concatenate all in one string.
# (this may not be really useful with recent Python versions) # (this may not be really useful with recent Python versions)
@ -660,7 +635,7 @@ class _OleStream(io.BytesIO):
# if size is zero, then first sector index should be ENDOFCHAIN: # if size is zero, then first sector index should be ENDOFCHAIN:
if size == 0 and sect != ENDOFCHAIN: if size == 0 and sect != ENDOFCHAIN:
debug('size == 0 and sect != ENDOFCHAIN:') debug('size == 0 and sect != ENDOFCHAIN:')
raise IOError, 'incorrect OLE sector index for empty stream' raise IOError('incorrect OLE sector index for empty stream')
#[PL] A fixed-length for loop is used instead of an undefined while #[PL] A fixed-length for loop is used instead of an undefined while
# loop to avoid DoS attacks: # loop to avoid DoS attacks:
for i in xrange(nb_sectors): for i in xrange(nb_sectors):
@ -671,17 +646,17 @@ class _OleStream(io.BytesIO):
else: else:
# else this means that the stream is smaller than declared: # else this means that the stream is smaller than declared:
debug('sect=ENDOFCHAIN before expected size') debug('sect=ENDOFCHAIN before expected size')
raise IOError, 'incomplete OLE stream' raise IOError('incomplete OLE stream')
# sector index should be within FAT: # sector index should be within FAT:
if sect<0 or sect>=len(fat): if sect<0 or sect>=len(fat):
debug('sect=%d (%X) / len(fat)=%d' % (sect, sect, len(fat))) debug('sect=%d (%X) / len(fat)=%d' % (sect, sect, len(fat)))
debug('i=%d / nb_sectors=%d' %(i, nb_sectors)) debug('i=%d / nb_sectors=%d' %(i, nb_sectors))
## tmp_data = string.join(data, "") ## tmp_data = b"".join(data)
## f = open('test_debug.bin', 'wb') ## f = open('test_debug.bin', 'wb')
## f.write(tmp_data) ## f.write(tmp_data)
## f.close() ## f.close()
## debug('data read so far: %d bytes' % len(tmp_data)) ## debug('data read so far: %d bytes' % len(tmp_data))
raise IOError, 'incorrect OLE FAT, sector index out of range' raise IOError('incorrect OLE FAT, sector index out of range')
#TODO: merge this code with OleFileIO.getsect() ? #TODO: merge this code with OleFileIO.getsect() ?
#TODO: check if this works with 4K sectors: #TODO: check if this works with 4K sectors:
try: try:
@ -689,7 +664,7 @@ class _OleStream(io.BytesIO):
except: except:
debug('sect=%d, seek=%d, filesize=%d' % debug('sect=%d, seek=%d, filesize=%d' %
(sect, offset+sectorsize*sect, filesize)) (sect, offset+sectorsize*sect, filesize))
raise IOError, 'OLE sector index out of range' raise IOError('OLE sector index out of range')
sector_data = fp.read(sectorsize) sector_data = fp.read(sectorsize)
# [PL] check if there was enough data: # [PL] check if there was enough data:
# Note: if sector is the last of the file, sometimes it is not a # Note: if sector is the last of the file, sometimes it is not a
@ -699,17 +674,17 @@ class _OleStream(io.BytesIO):
debug('sect=%d / len(fat)=%d, seek=%d / filesize=%d, len read=%d' % debug('sect=%d / len(fat)=%d, seek=%d / filesize=%d, len read=%d' %
(sect, len(fat), offset+sectorsize*sect, filesize, len(sector_data))) (sect, len(fat), offset+sectorsize*sect, filesize, len(sector_data)))
debug('seek+len(read)=%d' % (offset+sectorsize*sect+len(sector_data))) debug('seek+len(read)=%d' % (offset+sectorsize*sect+len(sector_data)))
raise IOError, 'incomplete OLE sector' raise IOError('incomplete OLE sector')
data.append(sector_data) data.append(sector_data)
# jump to next sector in the FAT: # jump to next sector in the FAT:
try: try:
sect = fat[sect] sect = fat[sect]
except IndexError: except IndexError:
# [PL] if pointer is out of the FAT an exception is raised # [PL] if pointer is out of the FAT an exception is raised
raise IOError, 'incorrect OLE FAT, sector index out of range' raise IOError('incorrect OLE FAT, sector index out of range')
#[PL] Last sector should be a "end of chain" marker: #[PL] Last sector should be a "end of chain" marker:
if sect != ENDOFCHAIN: if sect != ENDOFCHAIN:
raise IOError, 'incorrect last sector index in OLE stream' raise IOError('incorrect last sector index in OLE stream')
data = b"".join(data) data = b"".join(data)
# Data is truncated to the actual stream size: # Data is truncated to the actual stream size:
if len(data) >= size: if len(data) >= size:
@ -723,8 +698,8 @@ class _OleStream(io.BytesIO):
else: else:
# read data is less than expected: # read data is less than expected:
debug('len(data)=%d, size=%d' % (len(data), size)) debug('len(data)=%d, size=%d' % (len(data), size))
raise IOError, 'OLE stream size is less than declared' raise IOError('OLE stream size is less than declared')
# when all data is read in memory, StringIO constructor is called # when all data is read in memory, BytesIO constructor is called
io.BytesIO.__init__(self, data) io.BytesIO.__init__(self, data)
# Then the _OleStream object can be used as a read-only file object. # Then the _OleStream object can be used as a read-only file object.
@ -827,10 +802,10 @@ class _OleDirectoryEntry:
self.sid_right, self.sid_child)) self.sid_right, self.sid_child))
# sizeHigh is only used for 4K sectors, it should be zero for 512 bytes # sizeHigh is only used for 4K sectors, it should be zero for 512 bytes
# sectors, BUT apparently some implementations set it as 0xFFFFFFFFL, 1 # sectors, BUT apparently some implementations set it as 0xFFFFFFFF, 1
# or some other value so it cannot be raised as a defect in general: # or some other value so it cannot be raised as a defect in general:
if olefile.sectorsize == 512: if olefile.sectorsize == 512:
if sizeHigh != 0 and sizeHigh != 0xFFFFFFFFL: if sizeHigh != 0 and sizeHigh != 0xFFFFFFFF:
debug('sectorsize=%d, sizeLow=%d, sizeHigh=%d (%X)' % debug('sectorsize=%d, sizeLow=%d, sizeHigh=%d (%X)' %
(olefile.sectorsize, sizeLow, sizeHigh, sizeHigh)) (olefile.sectorsize, sizeLow, sizeHigh, sizeHigh))
olefile._raise_defect(DEFECT_UNSURE, 'incorrect OLE stream size') olefile._raise_defect(DEFECT_UNSURE, 'incorrect OLE stream size')
@ -906,7 +881,7 @@ class _OleDirectoryEntry:
self.append_kids(child.sid_left) self.append_kids(child.sid_left)
# Check if its name is not already used (case-insensitive): # Check if its name is not already used (case-insensitive):
name_lower = child.name.lower() name_lower = child.name.lower()
if self.kids_dict.has_key(name_lower): if name_lower in self.kids_dict:
self.olefile._raise_defect(DEFECT_INCORRECT, self.olefile._raise_defect(DEFECT_INCORRECT,
"Duplicate filename in OLE storage") "Duplicate filename in OLE storage")
# Then the child_sid _OleDirectoryEntry object is appended to the # Then the child_sid _OleDirectoryEntry object is appended to the
@ -1059,7 +1034,7 @@ class OleFileIO:
self.fp = filename self.fp = filename
else: else:
# string-like object: filename of file on disk # string-like object: filename of file on disk
#TODO: if larger than 1024 bytes, this could be the actual data => StringIO #TODO: if larger than 1024 bytes, this could be the actual data => BytesIO
self.fp = open(filename, "rb") self.fp = open(filename, "rb")
# old code fails if filename is not a plain string: # old code fails if filename is not a plain string:
#if isPath(filename): #if isPath(filename):
@ -1149,10 +1124,10 @@ class OleFileIO:
) = struct.unpack(fmt_header, header1) ) = struct.unpack(fmt_header, header1)
debug( struct.unpack(fmt_header, header1)) debug( struct.unpack(fmt_header, header1))
if self.Sig != '\xd0\xcf\x11\xe0\xa1\xb1\x1a\xe1': if self.Sig != b'\xd0\xcf\x11\xe0\xa1\xb1\x1a\xe1':
# OLE signature should always be present # OLE signature should always be present
self._raise_defect(DEFECT_FATAL, "incorrect OLE signature") self._raise_defect(DEFECT_FATAL, "incorrect OLE signature")
if self.clsid != '\x00'*16: if self.clsid != bytearray(16):
# according to AAF specs, CLSID should always be zero # according to AAF specs, CLSID should always be zero
self._raise_defect(DEFECT_INCORRECT, "incorrect CLSID in OLE header") self._raise_defect(DEFECT_INCORRECT, "incorrect CLSID in OLE header")
debug( "MinorVersion = %d" % self.MinorVersion ) debug( "MinorVersion = %d" % self.MinorVersion )
@ -1197,7 +1172,7 @@ class OleFileIO:
# calculate the number of sectors in the file # calculate the number of sectors in the file
# (-1 because header doesn't count) # (-1 because header doesn't count)
self.nb_sect = ( (filesize + self.SectorSize-1) / self.SectorSize) - 1 self.nb_sect = ( (filesize + self.SectorSize-1) // self.SectorSize) - 1
debug( "Number of sectors in the file: %d" % self.nb_sect ) debug( "Number of sectors in the file: %d" % self.nb_sect )
# file clsid (probably never used, so we don't store it) # file clsid (probably never used, so we don't store it)
@ -1271,14 +1246,14 @@ class OleFileIO:
DIFSECT: "DIFSECT " DIFSECT: "DIFSECT "
} }
nbsect = len(fat) nbsect = len(fat)
nlines = (nbsect+VPL-1)/VPL nlines = (nbsect+VPL-1)//VPL
print "index", print("index", end=" ")
for i in range(VPL): for i in range(VPL):
print ("%8X" % i), print("%8X" % i, end=" ")
print "" print()
for l in range(nlines): for l in range(nlines):
index = l*VPL index = l*VPL
print ("%8X:" % (firstindex+index)), print("%8X:" % (firstindex+index), end=" ")
for i in range(index, index+VPL): for i in range(index, index+VPL):
if i>=nbsect: if i>=nbsect:
break break
@ -1290,8 +1265,8 @@ class OleFileIO:
nom = " --->" nom = " --->"
else: else:
nom = "%8X" % sect nom = "%8X" % sect
print nom, print(nom, end=" ")
print "" print()
def dumpsect(self, sector, firstindex=0): def dumpsect(self, sector, firstindex=0):
@ -1301,21 +1276,21 @@ class OleFileIO:
VPL=8 # number of values per line (8+1 * 8+1 = 81) VPL=8 # number of values per line (8+1 * 8+1 = 81)
tab = array.array(UINT32, sector) tab = array.array(UINT32, sector)
nbsect = len(tab) nbsect = len(tab)
nlines = (nbsect+VPL-1)/VPL nlines = (nbsect+VPL-1)//VPL
print "index", print("index", end=" ")
for i in range(VPL): for i in range(VPL):
print ("%8X" % i), print("%8X" % i, end=" ")
print "" print()
for l in range(nlines): for l in range(nlines):
index = l*VPL index = l*VPL
print ("%8X:" % (firstindex+index)), print("%8X:" % (firstindex+index), end=" ")
for i in range(index, index+VPL): for i in range(index, index+VPL):
if i>=nbsect: if i>=nbsect:
break break
sect = tab[i] sect = tab[i]
nom = "%8X" % sect nom = "%8X" % sect
print nom, print(nom, end=" ")
print "" print()
def sect2array(self, sect): def sect2array(self, sect):
""" """
@ -1345,7 +1320,7 @@ class OleFileIO:
self.dumpsect(sect) self.dumpsect(sect)
# The FAT is a sector chain starting at the first index of itself. # The FAT is a sector chain starting at the first index of itself.
for isect in fat1: for isect in fat1:
#print "isect = %X" % isect #print("isect = %X" % isect)
if isect == ENDOFCHAIN or isect == FREESECT: if isect == ENDOFCHAIN or isect == FREESECT:
# the end of the sector chain has been reached # the end of the sector chain has been reached
break break
@ -1367,7 +1342,7 @@ class OleFileIO:
# described by DIF blocks # described by DIF blocks
sect = header[76:512] sect = header[76:512]
debug( "len(sect)=%d, so %d integers" % (len(sect), len(sect)/4) ) debug( "len(sect)=%d, so %d integers" % (len(sect), len(sect)//4) )
#fat = [] #fat = []
# [PL] FAT is an array of 32 bits unsigned ints, it's more effective # [PL] FAT is an array of 32 bits unsigned ints, it's more effective
# to use an array than a list in Python. # to use an array than a list in Python.
@ -1377,8 +1352,8 @@ class OleFileIO:
#self.dumpfat(self.fat) #self.dumpfat(self.fat)
## for i in range(0, len(sect), 4): ## for i in range(0, len(sect), 4):
## ix = i32(sect, i) ## ix = i32(sect, i)
## #[PL] if ix == -2 or ix == -1: # ix == 0xFFFFFFFEL or ix == 0xFFFFFFFFL: ## #[PL] if ix == -2 or ix == -1: # ix == 0xFFFFFFFE or ix == 0xFFFFFFFF:
## if ix == 0xFFFFFFFEL or ix == 0xFFFFFFFFL: ## if ix == 0xFFFFFFFE or ix == 0xFFFFFFFF:
## break ## break
## s = self.getsect(ix) ## s = self.getsect(ix)
## #fat = fat + [i32(s, i) for i in range(0, len(s), 4)] ## #fat = fat + [i32(s, i) for i in range(0, len(s), 4)]
@ -1396,10 +1371,10 @@ class OleFileIO:
debug( "DIFAT analysis..." ) debug( "DIFAT analysis..." )
# We compute the necessary number of DIFAT sectors : # We compute the necessary number of DIFAT sectors :
# (each DIFAT sector = 127 pointers + 1 towards next DIFAT sector) # (each DIFAT sector = 127 pointers + 1 towards next DIFAT sector)
nb_difat = (self.csectFat-109 + 126)/127 nb_difat = (self.csectFat-109 + 126)//127
debug( "nb_difat = %d" % nb_difat ) debug( "nb_difat = %d" % nb_difat )
if self.csectDif != nb_difat: if self.csectDif != nb_difat:
raise IOError, 'incorrect DIFAT' raise IOError('incorrect DIFAT')
isect_difat = self.sectDifStart isect_difat = self.sectDifStart
for i in xrange(nb_difat): for i in xrange(nb_difat):
debug( "DIFAT block %d, sector %X" % (i, isect_difat) ) debug( "DIFAT block %d, sector %X" % (i, isect_difat) )
@ -1414,11 +1389,11 @@ class OleFileIO:
# checks: # checks:
if isect_difat not in [ENDOFCHAIN, FREESECT]: if isect_difat not in [ENDOFCHAIN, FREESECT]:
# last DIFAT pointer value must be ENDOFCHAIN or FREESECT # last DIFAT pointer value must be ENDOFCHAIN or FREESECT
raise IOError, 'incorrect end of DIFAT' raise IOError('incorrect end of DIFAT')
## if len(self.fat) != self.csectFat: ## if len(self.fat) != self.csectFat:
## # FAT should contain csectFat blocks ## # FAT should contain csectFat blocks
## print "FAT length: %d instead of %d" % (len(self.fat), self.csectFat) ## print("FAT length: %d instead of %d" % (len(self.fat), self.csectFat))
## raise IOError, 'incorrect DIFAT' ## raise IOError('incorrect DIFAT')
# since FAT is read from fixed-size sectors, it may contain more values # since FAT is read from fixed-size sectors, it may contain more values
# than the actual number of sectors in the file. # than the actual number of sectors in the file.
# Keep only the relevant sector indexes: # Keep only the relevant sector indexes:
@ -1444,7 +1419,7 @@ class OleFileIO:
# 2) Actually used size is calculated by dividing the MiniStream size # 2) Actually used size is calculated by dividing the MiniStream size
# (given by root entry size) by the size of mini sectors, *4 for # (given by root entry size) by the size of mini sectors, *4 for
# 32 bits indexes: # 32 bits indexes:
nb_minisectors = (self.root.size + self.MiniSectorSize-1) / self.MiniSectorSize nb_minisectors = (self.root.size + self.MiniSectorSize-1) // self.MiniSectorSize
used_size = nb_minisectors * 4 used_size = nb_minisectors * 4
debug('loadminifat(): minifatsect=%d, nb FAT sectors=%d, used_size=%d, stream_size=%d, nb MiniSectors=%d' % debug('loadminifat(): minifatsect=%d, nb FAT sectors=%d, used_size=%d, stream_size=%d, nb MiniSectors=%d' %
(self.minifatsect, self.csectMiniFat, used_size, stream_size, nb_minisectors)) (self.minifatsect, self.csectMiniFat, used_size, stream_size, nb_minisectors))
@ -1473,7 +1448,7 @@ class OleFileIO:
# 512 bytes: # 512 bytes:
#self.fp.seek(512 + self.sectorsize * sect) #self.fp.seek(512 + self.sectorsize * sect)
#[PL]: added safety checks: #[PL]: added safety checks:
#print "getsect(%X)" % sect #print("getsect(%X)" % sect)
try: try:
self.fp.seek(self.sectorsize * (sect+1)) self.fp.seek(self.sectorsize * (sect+1))
except: except:
@ -1502,7 +1477,7 @@ class OleFileIO:
#[PL] to detect malformed documents and avoid DoS attacks, the maximum #[PL] to detect malformed documents and avoid DoS attacks, the maximum
# number of directory entries can be calculated: # number of directory entries can be calculated:
max_entries = self.directory_fp.size / 128 max_entries = self.directory_fp.size // 128
debug('loaddirectory: size=%d, max_entries=%d' % debug('loaddirectory: size=%d, max_entries=%d' %
(self.directory_fp.size, max_entries)) (self.directory_fp.size, max_entries))
@ -1659,7 +1634,7 @@ class OleFileIO:
def openstream(self, filename): def openstream(self, filename):
""" """
Open a stream as a read-only file object (StringIO). Open a stream as a read-only file object (BytesIO).
filename: path of stream in storage tree (except root entry), either: filename: path of stream in storage tree (except root entry), either:
- a string using Unix path syntax, for example: - a string using Unix path syntax, for example:
@ -1754,7 +1729,7 @@ class OleFileIO:
entry = self.direntries[sid] entry = self.direntries[sid]
if entry.entry_type != STGTY_STREAM: if entry.entry_type != STGTY_STREAM:
#TODO: Should it return zero instead of raising an exception ? #TODO: Should it return zero instead of raising an exception ?
raise TypeError, 'object is not an OLE stream' raise TypeError('object is not an OLE stream')
return entry.size return entry.size
@ -1799,17 +1774,16 @@ class OleFileIO:
fp.seek(i32(s, 16)) fp.seek(i32(s, 16))
# get section # get section
s = "****" + fp.read(i32(fp.read(4))-4) s = b"****" + fp.read(i32(fp.read(4))-4)
# number of properties: # number of properties:
num_props = i32(s, 4) num_props = i32(s, 4)
except: except BaseException as exc:
# catch exception while parsing property header, and only raise # catch exception while parsing property header, and only raise
# a DEFECT_INCORRECT then return an empty dict, because this is not # a DEFECT_INCORRECT then return an empty dict, because this is not
# a fatal error when parsing the whole file # a fatal error when parsing the whole file
exctype, excvalue = sys.exc_info()[:2]
msg = 'Error while parsing properties header in stream %s: %s' % ( msg = 'Error while parsing properties header in stream %s: %s' % (
repr(streampath), excvalue) repr(streampath), exc)
self._raise_defect(DEFECT_INCORRECT, msg, exctype) self._raise_defect(DEFECT_INCORRECT, msg, type(exc))
return data return data
for i in range(num_props): for i in range(num_props):
@ -1845,7 +1819,7 @@ class OleFileIO:
count = i32(s, offset+4) count = i32(s, offset+4)
value = s[offset+8:offset+8+count-1] value = s[offset+8:offset+8+count-1]
# remove all null chars: # remove all null chars:
value = value.replace('\x00', '') value = value.replace(b'\x00', b'')
elif type == VT_BLOB: elif type == VT_BLOB:
# binary large object (BLOB) # binary large object (BLOB)
# see http://msdn.microsoft.com/en-us/library/dd942282.aspx # see http://msdn.microsoft.com/en-us/library/dd942282.aspx
@ -1864,12 +1838,12 @@ class OleFileIO:
# since Jan 1,1601". # since Jan 1,1601".
if convert_time and id not in no_conversion: if convert_time and id not in no_conversion:
debug('Converting property #%d to python datetime, value=%d=%fs' debug('Converting property #%d to python datetime, value=%d=%fs'
%(id, value, float(value)/10000000L)) %(id, value, float(value)/10000000))
# convert FILETIME to Python datetime.datetime # convert FILETIME to Python datetime.datetime
# inspired from http://code.activestate.com/recipes/511425-filetime-to-datetime/ # inspired from http://code.activestate.com/recipes/511425-filetime-to-datetime/
_FILETIME_null_date = datetime.datetime(1601, 1, 1, 0, 0, 0) _FILETIME_null_date = datetime.datetime(1601, 1, 1, 0, 0, 0)
debug('timedelta days=%d' % (value/(10*1000000*3600*24))) debug('timedelta days=%d' % (value//(10*1000000*3600*24)))
value = _FILETIME_null_date + datetime.timedelta(microseconds=value/10) value = _FILETIME_null_date + datetime.timedelta(microseconds=value//10)
else: else:
# legacy code kept for backward compatibility: returns a # legacy code kept for backward compatibility: returns a
# number of seconds since Jan 1,1601 # number of seconds since Jan 1,1601
@ -1901,17 +1875,16 @@ class OleFileIO:
# type of items, e.g. VT_VECTOR|VT_BSTR # type of items, e.g. VT_VECTOR|VT_BSTR
# see http://msdn.microsoft.com/en-us/library/dd942011.aspx # see http://msdn.microsoft.com/en-us/library/dd942011.aspx
#print "%08x" % id, repr(value), #print("%08x" % id, repr(value), end=" ")
#print "(%s)" % VT[i32(s, offset) & 0xFFF] #print("(%s)" % VT[i32(s, offset) & 0xFFF])
data[id] = value data[id] = value
except: except BaseException as exc:
# catch exception while parsing each property, and only raise # catch exception while parsing each property, and only raise
# a DEFECT_INCORRECT, because parsing can go on # a DEFECT_INCORRECT, because parsing can go on
exctype, excvalue = sys.exc_info()[:2]
msg = 'Error while parsing property id %d in stream %s: %s' % ( msg = 'Error while parsing property id %d in stream %s: %s' % (
id, repr(streampath), excvalue) id, repr(streampath), exc)
self._raise_defect(DEFECT_INCORRECT, msg, exctype) self._raise_defect(DEFECT_INCORRECT, msg, type(exc))
return data return data
@ -1938,8 +1911,8 @@ if __name__ == "__main__":
# [PL] display quick usage info if launched from command-line # [PL] display quick usage info if launched from command-line
if len(sys.argv) <= 1: if len(sys.argv) <= 1:
print __doc__ print(__doc__)
print """ print("""
Launched from command line, this script parses OLE files and prints info. Launched from command line, this script parses OLE files and prints info.
Usage: OleFileIO_PL.py [-d] [-c] <file> [file2 ...] Usage: OleFileIO_PL.py [-d] [-c] <file> [file2 ...]
@ -1947,7 +1920,7 @@ Usage: OleFileIO_PL.py [-d] [-c] <file> [file2 ...]
Options: Options:
-d : debug mode (display a lot of debug information, for developers only) -d : debug mode (display a lot of debug information, for developers only)
-c : check all streams (for debugging purposes) -c : check all streams (for debugging purposes)
""" """)
sys.exit() sys.exit()
check_streams = False check_streams = False
@ -1975,65 +1948,66 @@ Options:
props = sorted(props.items()) props = sorted(props.items())
for k, v in props: for k, v in props:
#[PL]: avoid to display too large or binary values: #[PL]: avoid to display too large or binary values:
if isinstance(v, basestring): if isinstance(v, (basestring, bytes)):
if len(v) > 50: if len(v) > 50:
v = v[:50] v = v[:50]
if isinstance(v, bytes):
# quick and dirty binary check: # quick and dirty binary check:
for c in (1,2,3,4,5,6,7,11,12,14,15,16,17,18,19,20, for c in (1,2,3,4,5,6,7,11,12,14,15,16,17,18,19,20,
21,22,23,24,25,26,27,28,29,30,31): 21,22,23,24,25,26,27,28,29,30,31):
if chr(c) in v: if c in bytearray(v):
v = '(binary data)' v = '(binary data)'
break break
print(" ", k, v) print(" ", k, v)
if check_streams: if check_streams:
# Read all streams to check if there are errors: # Read all streams to check if there are errors:
print '\nChecking streams...' print('\nChecking streams...')
for streamname in ole.listdir(): for streamname in ole.listdir():
# print name using repr() to convert binary chars to \xNN: # print name using repr() to convert binary chars to \xNN:
print '-', repr('/'.join(streamname)),'-', print('-', repr('/'.join(streamname)),'-', end=' ')
st_type = ole.get_type(streamname) st_type = ole.get_type(streamname)
if st_type == STGTY_STREAM: if st_type == STGTY_STREAM:
print 'size %d' % ole.get_size(streamname) print('size %d' % ole.get_size(streamname))
# just try to read stream in memory: # just try to read stream in memory:
ole.openstream(streamname) ole.openstream(streamname)
else: else:
print 'NOT a stream : type=%d' % st_type print('NOT a stream : type=%d' % st_type)
print '' print()
## for streamname in ole.listdir(): ## for streamname in ole.listdir():
## # print name using repr() to convert binary chars to \xNN: ## # print name using repr() to convert binary chars to \xNN:
## print '-', repr('/'.join(streamname)),'-', ## print('-', repr('/'.join(streamname)),'-', end=' ')
## print ole.getmtime(streamname) ## print(ole.getmtime(streamname))
## print '' ## print()
print 'Modification/Creation times of all directory entries:' print('Modification/Creation times of all directory entries:')
for entry in ole.direntries: for entry in ole.direntries:
if entry is not None: if entry is not None:
print '- %s: mtime=%s ctime=%s' % (entry.name, print('- %s: mtime=%s ctime=%s' % (entry.name,
entry.getmtime(), entry.getctime()) entry.getmtime(), entry.getctime()))
print '' print()
# parse and display metadata: # parse and display metadata:
meta = ole.get_metadata() meta = ole.get_metadata()
meta.dump() meta.dump()
print '' print()
#[PL] Test a few new methods: #[PL] Test a few new methods:
root = ole.get_rootentry_name() root = ole.get_rootentry_name()
print 'Root entry name: "%s"' % root print('Root entry name: "%s"' % root)
if ole.exists('worddocument'): if ole.exists('worddocument'):
print "This is a Word document." print("This is a Word document.")
print "type of stream 'WordDocument':", ole.get_type('worddocument') print("type of stream 'WordDocument':", ole.get_type('worddocument'))
print "size :", ole.get_size('worddocument') print("size :", ole.get_size('worddocument'))
if ole.exists('macros/vba'): if ole.exists('macros/vba'):
print "This document may contain VBA macros." print("This document may contain VBA macros.")
# print parsing issues: # print parsing issues:
print '\nNon-fatal issues raised during parsing:' print('\nNon-fatal issues raised during parsing:')
if ole.parsing_issues: if ole.parsing_issues:
for exctype, msg in ole.parsing_issues: for exctype, msg in ole.parsing_issues:
print '- %s: %s' % (exctype.__name__, msg) print('- %s: %s' % (exctype.__name__, msg))
else: else:
print 'None' print('None')
## except IOError as v: ## except IOError as v:
## print("***", "cannot read", file, "-", v) ## print("***", "cannot read", file, "-", v)