Python ints passed into C++ functions as longs using ctypes get truncated

Question

I'm developing on a Mac using MacOSX 10.8.2 with the latest xcode and the stock python interpreter.

Here's some code I put into fails.cpp

#include <iostream>

using namespace std;

extern "C" {

void mysort(long *data, long data2) {
    cout << hex << *data << ' ' << data2 << endl;
}

}

Here's some code to call it from python that I put in fails.py:

import ctypes

sort_dll = ctypes.CDLL("fails.dylib")
mysort = ctypes.CFUNCTYPE(None, ctypes.POINTER(ctypes.c_long), ctypes.c_long)(sort_dll.mysort)

a = ctypes.c_long(0x987654321)
mysort(ctypes.byref(a), 0x123456789)

I compile and run with

c++ -arch x86_64 -o fails.o -c fails.cpp && g++ -o fails.dylib -dynamiclib fails.o && python fails.py

The result is:

987654321 23456789

Why is the 64-bit integer passed by value being truncated to 32 bits? Surprisingly, the pointer to a 64-bit long isn't truncated.

Josh Heitzman · Accepted Answer · 2012-11-16 02:42:04Z

1

I suspect it is because the Python ctypes library has decided that c_long is 32-bits in size. There are some hints in the docs:

Represents the C signed int datatype. The constructor accepts an optional integer initializer; no overflow checking is done. On platforms where sizeof(int) == sizeof(long) it is an alias to c_long.

This code will tell you how big c_long actually is:

import ctypes
print ctypes.sizeof(ctypes.c_long())

The value the pointer references wouldn't be truncated since ctypes only has to marshall the pointer itself. It's possible the pointer is being truncated, but it doesn't matter as the high bits are all zero anyway. It's also possible that ctypes decided that ctypes.POINTER is 64-bit. You can find out by modifying the above example just slightly.

answered Nov 16, 2012 at 2:42

Josh Heitzman

1,8341 gold badge14 silver badges26 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Dan Piponi Over a year ago

Good idea. But ctypes.sizeof(ctypes.c_long()) gives 8 so it's not truncating because it has the wrong size for c_long. I'm inclined to think it's a bug though I'd have though such a blatant bug would have been caught by basic testing.

Josh Heitzman Over a year ago

It's possible the bug is in the implicit conversion from a Python int to a c_long. As a workaround, it is probably worth trying the explicit ctypes.c_long(0x123456789) as a workaround or even the int64.

Dan Piponi Over a year ago

I've tried quite a few combinations of code explicitly converting to 64-bit ints and they all seem to fail. My workaround is to pass by reference. It works fine but I don't know what I can trust...

Collectives™ on Stack Overflow

Python ints passed into C++ functions as longs using ctypes get truncated

1 Answer 1

3 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Related