structure.py

# Copyright (c) 2003-2012 CORE Security Technologies
#
# This software is provided under under a slightly modified version
# of the Apache Software License. See the accompanying LICENSE file
# for more information.
#
# $Id$
#
from __future__ import print_function
from struct import pack, unpack, calcsize

debug = 0

class Structure:
	""" sublcasses can define commonHdr and/or structure.
		each of them is an tuple of either two: (fieldName, format) or three: (fieldName, ':', class) fields.
		[it can't be a dictionary, because order is important]
		
		where format specifies how the data in the field will be converted to/from bytes (string)
		class is the class to use when unpacking ':' fields.

		each field can only contain one value (or an array of values for *)
			i.e. struct.pack('Hl',1,2) is valid, but format specifier 'Hl' is not (you must use 2 dfferent fields)

		format specifiers:
			specifiers from module pack can be used with the same format 
			see struct.__doc__ (pack/unpack is finally called)
				x       [padding byte]
				c       [character]
				b       [signed byte]
				B       [unsigned byte]
				h       [signed short]
				H       [unsigned short]
				l       [signed long]
				L       [unsigned long]
				i       [signed integer]
				I       [unsigned integer]
				q       [signed long long (quad)]
				Q       [unsigned long long (quad)]
				s       [string (array of chars), must be preceded with length in format specifier, padded with zeros]
				p       [pascal string (includes byte count), must be preceded with length in format specifier, padded with zeros]
				f       [float]
				d       [double]
				=       [native byte ordering, size and alignment]
				@       [native byte ordering, standard size and alignment]
				!       [network byte ordering]
				<       [little endian]
				>       [big endian]

			usual printf like specifiers can be used (if started with %) 
			[not recommeneded, there is no why to unpack this]

				%08x    will output an 8 bytes hex
				%s      will output a string
				%s\\x00  will output a NUL terminated string
				%d%d    will output 2 decimal digits (against the very same specification of Structure)
				...

			some additional format specifiers:
				:       just copy the bytes from the field into the output string (input may be string, other structure, or anything responding to __str__()) (for unpacking, all what's left is returned)
				z       same as :, but adds a NUL byte at the end (asciiz) (for unpacking the first NUL byte is used as terminator)  [asciiz string]
				u       same as z, but adds two NUL bytes at the end (after padding to an even size with NULs). (same for unpacking) [unicode string]
				w       DCE-RPC/NDR string (it's a macro for [  '<L=(len(field)+1)/2','"\\x00\\x00\\x00\\x00','<L=(len(field)+1)/2',':' ]
				?-field length of field named 'field', formated as specified with ? ('?' may be '!H' for example). The input value overrides the real length
				?1*?2   array of elements. Each formated as '?2', the number of elements in the array is stored as specified by '?1' (?1 is optional, or can also be a constant (number), for unpacking)
				'xxxx   literal xxxx (field's value doesn't change the output. quotes must not be closed or escaped)
				"xxxx   literal xxxx (field's value doesn't change the output. quotes must not be closed or escaped)
				_       will not pack the field. Accepts a third argument, which is an unpack code. See _Test_UnpackCode for an example
				?=packcode  will evaluate packcode in the context of the structure, and pack the result as specified by ?. Unpacking is made plain
				?&fieldname "Address of field fieldname".
						For packing it will simply pack the id() of fieldname. Or use 0 if fieldname doesn't exists.
						For unpacking, it's used to know weather fieldname has to be unpacked or not, i.e. by adding a & field you turn another field (fieldname) in an optional field.

	"""
	commonHdr = ()
	structure = ()

	def __init__(self, data = None, alignment = 0):
		if not hasattr(self, 'alignment'):
			self.alignment = alignment

		self.fields    = {}
		if data is not None:
			self.fromString(data)
		else:
			self.data = None

	def packField(self, fieldName, format = None):
		if debug:
			print("packField( %s | %s )" % (fieldName, format))

		if format is None:
			format = self.formatForField(fieldName)

		if fieldName in self.fields:
			ans = self.pack(format, self.fields[fieldName], field = fieldName)
		else:
			ans = self.pack(format, None, field = fieldName)

		if debug:
			print("\tanswer %r" % ans)

		return ans

	def getData(self):
		if self.data is not None:
			return self.data
		data = b''
		for field in self.commonHdr+self.structure:
			try:
				data += self.packField(field[0], field[1])
			except Exception as e:
				if field[0] in self.fields:
					e.args += ("When packing field '%s | %s | %r' in %s" % (field[0], field[1], self[field[0]], self.__class__),)
				else:
					e.args += ("When packing field '%s | %s' in %s" % (field[0], field[1], self.__class__),)
				raise
			if self.alignment:
				if len(data) % self.alignment:
					data += (b'\x00'*self.alignment)[:-(len(data) % self.alignment)]
			
		#if len(data) % self.alignment: data += ('\x00'*self.alignment)[:-(len(data) % self.alignment)]
		return data

	def fromString(self, data):
		for field in self.commonHdr+self.structure:
			if debug:
				print("fromString( %s | %s | %r )" % (field[0], field[1], data))
			size = self.calcUnpackSize(field[1], data, field[0])
			if debug:
				print("  size = %d" % size)
			dataClassOrCode = bytes
			if len(field) > 2:
				dataClassOrCode = field[2]
			try:
				self[field[0]] = self.unpack(field[1], data[:size], dataClassOrCode = dataClassOrCode, field = field[0])
			except Exception as e:
				e.args += ("When unpacking field '%s | %s | %r[:%d]'" % (field[0], field[1], data, size),)
				raise

			size = self.calcPackSize(field[1], self[field[0]], field[0])
			if self.alignment and size % self.alignment:
				size += self.alignment - (size % self.alignment)
			data = data[size:]

		if debug:
			self.dump(indent=4)
		return self
		
	def __setitem__(self, key, value):
		self.fields[key] = value
		self.data = None        # force recompute

	def __getitem__(self, key):
		return self.fields[key]

	def __delitem__(self, key):
		del self.fields[key]
		
	def __bytes__(self):
		if debug:
			self.dump(indent=4)
		return self.getData()

	def __str__(self):
		"""
		In python 2, func `bytes` is alias of `str` and redirect to `__str__`,
		workaround here is to redirect back to `__bytes__`
		"""
		if str is bytes:
			return self.__bytes__()
		else:
			return super(Structure, self).__str__()

	def __len__(self):
		# XXX: improve
		return len(self.getData())

	def pack(self, format, data, field = None):
		if debug:
			print("  pack( %s | %r | %s)" %  (format, data, field))

		if field:
			addressField = self.findAddressFieldFor(field)
			if (addressField is not None) and (data is None):
				return b''

		# void specifier
		if format[:1] == '_':
			return b''

		# quote specifier
		if format[:1] == "'" or format[:1] == '"':
			return format[1:].encode()

		# code specifier
		two = format.split('=')
		if len(two) >= 2:
			if data:
				return self.pack(two[0], data)
			else:
				fields = {'self':self}
				fields.update(self.fields)
				self[field] = eval(two[1], {}, fields)
				return self.pack(two[0], self[field])

		# address specifier
		two = format.split('&')
		if len(two) == 2:
			if data:
				return self.pack(two[0], data)
			else:
				if (two[1] in self.fields) and (self[two[1]] is not None):
					self[field] = id(self[two[1]]) & ((1<<(calcsize(two[0])*8))-1)
				else:
					self[field] = 0
				return self.pack(two[0], self[field])

		# length specifier
		two = format.split('-')
		if len(two) == 2:
			if data:
				return self.pack(two[0],data)
			else:
				self[field] = self.calcPackFieldSize(two[1])
				return self.pack(two[0], self[field])

		# array specifier
		two = format.split('*')
		if len(two) == 2:
			answer = b''
			for each in data:
				answer += self.pack(two[1], each)
			if two[0]:
				if two[0].isdigit():
					if int(two[0]) != len(data):
						raise Exception("Array field has a constant size, and it doesn't match the actual value")
				else:
					return self.pack(two[0], len(data))+answer
			return answer

		# "printf" string specifier
		if format[:1] == '%':
			# format string like specifier
			return format % data

		# asciiz specifier
		if format[:1] == 'z':
			return data+b'\0'

		# unicode specifier
		if format[:1] == 'u':
			return data+b'\0\0' + (len(data) & 1 and b'\0' or b'')

		# DCE-RPC/NDR string specifier
		if format[:1] == 'w':
			if len(data) == 0:
				data = b'\0\0'
			elif len(data) % 2:
				data = data.encode() + b'\0'
			l = pack('<L', len(data)//2)
			return b''.join(map(bytes,[l, b'\0\0\0\0', l, data]))
					
		if data is None:
			raise Exception("Trying to pack None")
		
		# literal specifier
		if format[:1] == ':':
			return bytes(data)

		# struct like specifier
		return pack(format, data)

	def unpack(self, format, data, dataClassOrCode = bytes, field = None):
		if debug:
			print("  unpack( %s | %r )" %  (format, data))

		if field:
			addressField = self.findAddressFieldFor(field)
			if addressField is not None:
				if not self[addressField]:
					return

		# void specifier
		if format[:1] == '_':
			if dataClassOrCode != bytes:
				fields = {'self':self, 'inputDataLeft':data}
				fields.update(self.fields)
				return eval(dataClassOrCode, {}, fields)
			else:
				return None

		# quote specifier
		if format[:1] == "'" or format[:1] == '"':
			answer = format[1:]
			if answer != data.decode():
				raise Exception("Unpacked data doesn't match constant value '%r' should be '%r'" % (data, answer))
			return answer

		# address specifier
		two = format.split('&')
		if len(two) == 2:
			return self.unpack(two[0],data)

		# code specifier
		two = format.split('=')
		if len(two) >= 2:
			return self.unpack(two[0],data)

		# length specifier
		two = format.split('-')
		if len(two) == 2:
			return self.unpack(two[0],data)

		# array specifier
		two = format.split('*')
		if len(two) == 2:
			answer = []
			sofar = 0
			if two[0].isdigit():
				number = int(two[0])
			elif two[0]:
				sofar += self.calcUnpackSize(two[0], data)
				number = self.unpack(two[0], data[:sofar])
			else:
				number = -1

			while number and sofar < len(data):
				nsofar = sofar + self.calcUnpackSize(two[1],data[sofar:])
				answer.append(self.unpack(two[1], data[sofar:nsofar], dataClassOrCode))
				number -= 1
				sofar = nsofar
			return answer

		# "printf" string specifier
		if format[:1] == '%':
			# format string like specifier
			return format % data

		# asciiz specifier
		if format == 'z':
			if data[-1:] != b'\x00':
				raise Exception("%s 'z' field is not NUL terminated: %r" % (field, data))
			return data[:-1] # remove trailing NUL

		# unicode specifier
		if format == 'u':
			if data[-2:] != b'\x00\x00':
				raise Exception("%s 'u' field is not NUL-NUL terminated: %r" % (field, data))
			return data[:-2] # remove trailing NUL

		# DCE-RPC/NDR string specifier
		if format == 'w':
			l = unpack('<L', data[:4])[0]
			return data[12:12+l*2]

		# literal specifier
		if format == ':':
			return dataClassOrCode(data)

		# struct like specifier
		return unpack(format, data)[0]

	def calcPackSize(self, format, data, field = None):
#        # print "  calcPackSize  %s:%r" %  (format, data)
		if field:
			addressField = self.findAddressFieldFor(field)
			if addressField is not None:
				if not self[addressField]:
					return 0

		# void specifier
		if format[:1] == '_':
			return 0

		# quote specifier
		if format[:1] == "'" or format[:1] == '"':
			return len(format)-1

		# address specifier
		two = format.split('&')
		if len(two) == 2:
			return self.calcPackSize(two[0], data)

		# code specifier
		two = format.split('=')
		if len(two) >= 2:
			return self.calcPackSize(two[0], data)

		# length specifier
		two = format.split('-')
		if len(two) == 2:
			return self.calcPackSize(two[0], data)

		# array specifier
		two = format.split('*')
		if len(two) == 2:
			answer = 0
			if two[0].isdigit():
					if int(two[0]) != len(data):
						raise Exception("Array field has a constant size, and it doesn't match the actual value")
			elif two[0]:
				answer += self.calcPackSize(two[0], len(data))

			for each in data:
				answer += self.calcPackSize(two[1], each)
			return answer

		# "printf" string specifier
		if format[:1] == '%':
			# format string like specifier
			return len(format % data)

		# asciiz specifier
		if format[:1] == 'z':
			return len(data)+1

		# asciiz specifier
		if format[:1] == 'u':
			l = len(data)
			return l + (l & 1 and 3 or 2)

		# DCE-RPC/NDR string specifier
		if format[:1] == 'w':
			l = len(data)
			return 12+l+l % 2

		# literal specifier
		if format[:1] == ':':
			return len(data)

		# struct like specifier
		return calcsize(format)

	def calcUnpackSize(self, format, data, field = None):
		if debug:
			print("  calcUnpackSize( %s | %s | %r)" %  (field, format, data))

		# void specifier
		if format[:1] == '_':
			return 0

		addressField = self.findAddressFieldFor(field)
		if addressField is not None:
			if not self[addressField]:
				return 0

		try:
			lengthField = self.findLengthFieldFor(field)
			return self[lengthField]
		except:
			pass

		# XXX: Try to match to actual values, raise if no match
		
		# quote specifier
		if format[:1] == "'" or format[:1] == '"':
			return len(format)-1

		# address specifier
		two = format.split('&')
		if len(two) == 2:
			return self.calcUnpackSize(two[0], data)

		# code specifier
		two = format.split('=')
		if len(two) >= 2:
			return self.calcUnpackSize(two[0], data)

		# length specifier
		two = format.split('-')
		if len(two) == 2:
			return self.calcUnpackSize(two[0], data)

		# array specifier
		two = format.split('*')
		if len(two) == 2:
			answer = 0
			if two[0]:
				if two[0].isdigit():
					number = int(two[0])
				else:
					answer += self.calcUnpackSize(two[0], data)
					number = self.unpack(two[0], data[:answer])

				while number:
					number -= 1
					answer += self.calcUnpackSize(two[1], data[answer:])
			else:
				while answer < len(data):
					answer += self.calcUnpackSize(two[1], data[answer:])
			return answer

		# "printf" string specifier
		if format[:1] == '%':
			raise Exception("Can't guess the size of a printf like specifier for unpacking")

		# asciiz specifier
		if format[:1] == 'z':
			return data.index(b'\x00')+1

		# asciiz specifier
		if format[:1] == 'u':
			l = data.index(b'\x00\x00')
			return l + (l & 1 and 3 or 2)

		# DCE-RPC/NDR string specifier
		if format[:1] == 'w':
			l = unpack('<L', data[:4])[0]
			return 12+l*2

		# literal specifier
		if format[:1] == ':':
			return len(data)

		# struct like specifier
		return calcsize(format)

	def calcPackFieldSize(self, fieldName, format = None):
		if format is None:
			format = self.formatForField(fieldName)

		return self.calcPackSize(format, self[fieldName])

	def formatForField(self, fieldName):
		for field in self.commonHdr+self.structure:
			if field[0] == fieldName:
				return field[1]
		raise Exception("Field %s not found" % fieldName)

	def findAddressFieldFor(self, fieldName):
		descriptor = '&%s' % fieldName
		l = len(descriptor)
		for field in self.commonHdr+self.structure:
			if field[1][-l:] == descriptor:
				return field[0]
		return None
		
	def findLengthFieldFor(self, fieldName):
		descriptor = '-%s' % fieldName
		l = len(descriptor)
		for field in self.commonHdr+self.structure:
			if field[1][-l:] == descriptor:
				return field[0]
		return None

	def dump(self, msg = None, indent = 0):
		if msg is None: msg = self.__class__.__name__
		ind = ' '*indent
		print("\n%s" % (msg))
		fixedFields = []
		for field in self.commonHdr+self.structure:
			i = field[0] 
			if i in self.fields:
				fixedFields.append(i)
				if isinstance(self[i], Structure):
					self[i].dump('%s%s:{' % (ind,i), indent = indent + 4)
					print("%s}" % ind)
				else:
					print("%s%s: {%r}" % (ind,i,self[i]))
		# Do we have remaining fields not defined in the structures? let's 
		# print them
		remainingFields = list(set(self.fields) - set(fixedFields))
		for i in remainingFields:
			if isinstance(self[i], Structure):
				self[i].dump('%s%s:{' % (ind,i), indent = indent + 4)
				print("%s}" % ind)
			else:
				print("%s%s: {%r}" % (ind,i,self[i]))


class _StructureTest:
	alignment = 0
	def create(self,data = None):
		if data is not None:
			return self.theClass(data, alignment = self.alignment)
		else:
			return self.theClass(alignment = self.alignment)

	def run(self):
		print()
		print("-"*70)
		testName = self.__class__.__name__
		print("starting test: %s....." % testName)
		a = self.create()
		self.populate(a)
		a.dump("packing.....")
		a_str = bytes(a)
		print("packed: %r" % a_str)
		print("unpacking.....")
		b = self.create(a_str)
		b.dump("unpacked.....")
		print("repacking.....")
		b_str = bytes(b)
		if b_str != a_str:
			print("ERROR: original packed and repacked don't match")
			print("packed: %r" % b_str)

class _Test_simple(_StructureTest):
	class theClass(Structure):
		commonHdr = ()
		structure = (
			('int1', '!L'),
			('len1','!L-z1'),
			('arr1','B*<L'),
			('z1', 'z'),
			('u1','u'),
			('', '"COCA'),
			('len2','!H-:1'),
			('', '"COCA'),
			(':1', ':'),
			('int3','>L'),
			('code1','>L=len(arr1)*2+0x1000'),
		)

	def populate(self, a):
		a['default'] = 'hola'
		a['int1'] = 0x3131
		a['int3'] = 0x45444342
		a['z1']   = b'hola'
		a['u1']   = 'hola'.encode('utf_16_le')
		a[':1']   = b':1234:'
		a['arr1'] = (0x12341234,0x88990077,0x41414141)
		# a['len1'] = 0x42424242

class _Test_fixedLength(_Test_simple):
	def populate(self, a):
		_Test_simple.populate(self, a)
		a['len1'] = 0x42424242

class _Test_simple_aligned4(_Test_simple):
	alignment = 4

class _Test_nested(_StructureTest):
	class theClass(Structure):
		class _Inner(Structure):
			structure = (('data', 'z'),)

		structure = (
			('nest1', ':', _Inner),
			('nest2', ':', _Inner),
			('int', '<L'),
		)

	def populate(self, a):
		a['nest1'] = _Test_nested.theClass._Inner()
		a['nest2'] = _Test_nested.theClass._Inner()
		a['nest1']['data'] = b'hola manola'
		a['nest2']['data'] = b'chau loco'
		a['int'] = 0x12345678
	
class _Test_Optional(_StructureTest):
	class theClass(Structure):
		structure = (
			('pName','<L&Name'),
			('pList','<L&List'),
			('Name','w'),
			('List','<H*<L'),
		)
			
	def populate(self, a):
		a['Name'] = 'Optional test'
		a['List'] = (1,2,3,4)
		
class _Test_Optional_sparse(_Test_Optional):
	def populate(self, a):
		_Test_Optional.populate(self, a)
		del a['Name']

class _Test_AsciiZArray(_StructureTest):
	class theClass(Structure):
		structure = (
			('head','<L'),
			('array','B*z'),
			('tail','<L'),
		)

	def populate(self, a):
		a['head'] = 0x1234
		a['tail'] = 0xabcd
		a['array'] = (b'hola',b'manola',b'te traje')
		
class _Test_UnpackCode(_StructureTest):
	class theClass(Structure):
		structure = (
			('leni','<L=len(uno)*2'),
			('cuchi','_-uno','leni//2'),
			('uno',':'),
			('dos',':'),
		)

	def populate(self, a):
		a['uno'] = b'soy un loco!'
		a['dos'] = b'que haces fiera'

class _Test_AAA(_StructureTest):
	class theClass(Structure):
		commonHdr = ()
		structure = (
			('iv', '!L=((init_vector & 0xFFFFFF) << 8) | ((pad & 0x3f) << 2) | (keyid & 3)'),
			('init_vector',   '_','(iv >> 8)'),
			('pad',           '_','((iv >>2) & 0x3F)'),
			('keyid',         '_','( iv & 0x03 )'),
			('dataLen',       '_-data', 'len(inputDataLeft)-4'),
			('data',':'),
			('icv','>L'),
		)

	def populate(self, a):
		a['init_vector']=0x01020304
		#a['pad']=int('01010101',2)
		a['pad']=int('010101',2)
		a['keyid']=0x07
		a['data']=b"\xA0\xA1\xA2\xA3\xA4\xA5\xA6\xA7\xA8\xA9"
		a['icv'] = 0x05060708
		#a['iv'] = 0x01020304
		
if __name__ == '__main__':
	_Test_simple().run()

	try:
		_Test_fixedLength().run()
	except:
		print("cannot repack because length is bogus")

	_Test_simple_aligned4().run()
	_Test_nested().run()
	_Test_Optional().run()
	_Test_Optional_sparse().run()
	_Test_AsciiZArray().run()
	_Test_UnpackCode().run()
	_Test_AAA().run()