Structure encoding feature (#1958)

* Structure: Adds a class variable "ENCODING" which specifies which encoding to use in case that a unicode string needs to be encoded (while packing/unpacking), defaults to latin-1 to keep compatibility with older versions

* Changed AsciiOrUnicodeStructure encoding from latin-1 to utf-8, fixes issue with share names containing non latin-1 characters.

* Re-added whitespaces
Added comment linking to PR and a brief explanation

* Re added whitespaces

Author: alexisbalbachan

impacket/smb.py

@@ -814,6 +814,8 @@ def __init__(self, commandOrData = None, data = None, **kargs):
 class AsciiOrUnicodeStructure(Structure):
     UnicodeStructure = ()
     AsciiStructure   = ()
+    ENCODING = 'utf-8'
+
     def __init__(self, flags = 0, **kargs):
         if flags & SMB.FLAGS2_UNICODE:
             self.structure = self.UnicodeStructure

impacket/structure.py

@@ -12,6 +12,8 @@
 from __future__ import division
 from __future__ import print_function
 from struct import pack, unpack, calcsize
+
+import six
 from six import b, PY3
 from binascii import hexlify
 
@@ -80,13 +82,19 @@ class is the class to use when unpacking ':' fields.
     commonHdr = ()
     structure = ()
     debug = 0
+    # Encoding defaults to latin-1 which already was the de facto encoding for structures and works for most use cases.
+    # Now it can be configured to another encoding if needed.
+    ENCODING = 'latin-1'   # https://github.com/fortra/impacket/pull/1958
 
     def __init__(self, data = None, alignment = 0):
         if not hasattr(self, 'alignment'):
             self.alignment = alignment
 
         self.fields    = {}
         self.rawData   = data
+
+        self.b = lambda x: six.ensure_binary(x, encoding=self.ENCODING)
+
         if data is not None:
             self.fromString(data)
         else:
@@ -197,7 +205,7 @@ def pack(self, format, data, field = None):
 
         # quote specifier
         if format[:1] == "'" or format[:1] == '"':
-            return b(format[1:])
+            return self.b(format[1:])
 
         # code specifier
         two = format.split('=')
@@ -245,26 +253,26 @@ def pack(self, format, data, field = None):
         # "printf" string specifier
         if format[:1] == '%':
             # format string like specifier
-            return b(format % data)
+            return self.b(format % data)
 
         # asciiz specifier
         if format[:1] == 'z':
             if isinstance(data,bytes):
-                return data + b('\0')
-            return bytes(b(data)+b('\0'))
+                return data + self.b('\0')
+            return bytes(self.b(data)+self.b('\0'))
 
         # unicode specifier
         if format[:1] == 'u':
-            return bytes(data+b('\0\0') + (len(data) & 1 and b('\0') or b''))
+            return bytes(data+self.b('\0\0') + (len(data) & 1 and self.b('\0') or b''))
 
         # DCE-RPC/NDR string specifier
         if format[:1] == 'w':
             if len(data) == 0:
-                data = b('\0\0')
+                data = self.b('\0\0')
             elif len(data) % 2:
-                data = b(data) + b('\0')
+                data = self.b(data) + self.b('\0')
             l = pack('<L', len(data)//2)
-            return b''.join([l, l, b('\0\0\0\0'), data])
+            return b''.join([l, l, self.b('\0\0\0\0'), data])
 
         if data is None:
             raise Exception("Trying to pack None")
@@ -279,7 +287,7 @@ def pack(self, format, data, field = None):
             elif isinstance(data, int):
                 return bytes(data)
             elif isinstance(data, bytes) is not True:
-                return bytes(b(data))
+                return bytes(self.b(data))
             else:
                 return data
 
@@ -288,7 +296,7 @@ def pack(self, format, data, field = None):
             if isinstance(data, bytes) or isinstance(data, bytearray):
                 return pack(format, data)
             else:
-                return pack(format, b(data))
+                return pack(format, self.b(data))
 
         # struct like specifier
         return pack(format, data)
@@ -315,7 +323,7 @@ def unpack(self, format, data, dataClassOrCode = b, field = None):
         # quote specifier
         if format[:1] == "'" or format[:1] == '"':
             answer = format[1:]
-            if b(answer) != data:
+            if self.b(answer) != data:
                 raise Exception("Unpacked data doesn't match constant value '%r' should be '%r'" % (data, answer))
             return answer
 
@@ -361,7 +369,7 @@ def unpack(self, format, data, dataClassOrCode = b, field = None):
 
         # asciiz specifier
         if format == 'z':
-            if data[-1:] != b('\x00'):
+            if data[-1:] != self.b('\x00'):
                 raise Exception("%s 'z' field is not NUL terminated: %r" % (field, data))
             if PY3:
                 return data[:-1].decode('latin-1')
@@ -370,7 +378,7 @@ def unpack(self, format, data, dataClassOrCode = b, field = None):
 
         # unicode specifier
         if format == 'u':
-            if data[-2:] != b('\x00\x00'):
+            if data[-2:] != self.b('\x00\x00'):
                 raise Exception("%s 'u' field is not NUL-NUL terminated: %r" % (field, data))
             return data[:-2] # remove trailing NUL
 
@@ -524,11 +532,11 @@ def calcUnpackSize(self, format, data, field = None):
 
         # asciiz specifier
         if format[:1] == 'z':
-            return data.index(b('\x00'))+1
+            return data.index(self.b('\x00'))+1
 
         # asciiz specifier
         if format[:1] == 'u':
-            l = data.index(b('\x00\x00'))
+            l = data.index(self.b('\x00\x00'))
             return l + (l & 1 and 3 or 2)
 
         # DCE-RPC/NDR string specifier
@@ -584,7 +592,7 @@ def zeroValue(self, format):
         if format in ['z',':','u']:
             return b''
         if format == 'w':
-            return b('\x00\x00')
+            return self.b('\x00\x00')
 
         return 0