I have need create a ctypes struct using a pointer. Example code

class MessageDecoder:
    def __init__(self, raw_data : bytes):
        self.raw_data = raw_data
        self.writable_raw_data = None
        self.datagram = None
        self.tag = None

    def decode(self):
            c_array = cast(self.raw_data, POINTER(c_ubyte * 
                      len(self.raw_data)))[0]
           data_pointer = pointer(c_array) 
            self.datagram = Datagram.from_buffer(data_pointer)

I get the c_array. I can get the pointer. However, trying to create Datagram fails. I can create one using c_array. EX: Datagram.from_buffer(c_array) Datagram looks like:

class Datagram(BigEndianStructure):
    _pack_ = 1
    _fields_ = [
        ("MagicNumber", c_uint),
        ("VersionMinor", c_uint16),
        ("VersionMajor", c_uint16),
        ("SequenceNumber", c_uint16),
        ("SeqNumberHigh", c_uint16),
        ("DataSize", c_uint32),
        ("Attribute", GenericAttribute)
    ]

Why?

I have need create a ctypes struct using a pointer. Example code

class MessageDecoder:
    def __init__(self, raw_data : bytes):
        self.raw_data = raw_data
        self.writable_raw_data = None
        self.datagram = None
        self.tag = None

    def decode(self):
            c_array = cast(self.raw_data, POINTER(c_ubyte * 
                      len(self.raw_data)))[0]
           data_pointer = pointer(c_array) 
            self.datagram = Datagram.from_buffer(data_pointer)

I get the c_array. I can get the pointer. However, trying to create Datagram fails. I can create one using c_array. EX: Datagram.from_buffer(c_array) Datagram looks like:

class Datagram(BigEndianStructure):
    _pack_ = 1
    _fields_ = [
        ("MagicNumber", c_uint),
        ("VersionMinor", c_uint16),
        ("VersionMajor", c_uint16),
        ("SequenceNumber", c_uint16),
        ("SeqNumberHigh", c_uint16),
        ("DataSize", c_uint32),
        ("Attribute", GenericAttribute)
    ]

Why?

• python
• ctypes

1 Answer 1

Much of decode is unnecessary. Use Datagram.from_buffer_copy. Copy is needed because the raw data is immutable bytes. If bytearray was used, then from_buffer could be used but then the datagram would share its internal data with the byte array.

Working example (dropped the undefined GenericAttribute):

import ctypes as ct
import struct

class Datagram(ct.BigEndianStructure):
    _pack_ = 1
    _fields_ = (("MagicNumber", ct.c_uint),
                ("VersionMinor", ct.c_uint16),
                ("VersionMajor", ct.c_uint16),
                ("SequenceNumber", ct.c_uint16),
                ("SeqNumberHigh", ct.c_uint16),
                ("DataSize", ct.c_uint32))

    def __repr__(self):
        return (f'Datagram(MagicNumber={self.MagicNumber}, '
                         f'VersionMinor={self.VersionMinor}, '
                         f'VersionMajor={self.VersionMajor}, '
                         f'SequenceNumber={self.SequenceNumber}, '
                         f'SeqNumberHigh={self.SeqNumberHigh}, '
                         f'DataSize={self.DataSize})')

raw_data = struct.pack('>LHHHHL',1,2,3,4,5,6)
print(raw_data.hex(' '))
datagram = Datagram.from_buffer_copy(raw_data)
print(datagram)

Output:

00 00 00 01 00 02 00 03 00 04 00 05 00 00 00 06
Datagram(MagicNumber=1, VersionMinor=2, VersionMajor=3, SequenceNumber=4, SeqNumberHigh=5, DataSize=6)