Question

Pointers are not common in Python, neither the var declaration type. I’m triyng to do this in Python.

C++:

 ...a function
 flt32 f;
 int32 sign, exp, man, *tms;
 tms = (int *) &f;
 ...(operations)
 *tms = sign | exp | man;
 return (f);

as you can see, tms points to f data, but just the int part of the float (btw, this function works perfectly).

With ctypes I can use pointers in Python:

from ctypes import * 
f = c_float(12.3)
tms = pointer(f) # tms should be: tms = c_int32(value)
print tms.contents.value #12.3000001907

The problem here is that tms becomes instantly the type of f, so both variables are float (tms should be tms = c_int32(a_value)).
It is possible to use a pointer that just matches the int value from a float variable in Python?

Answer 1

If you really want to do this in Python, you can use ctypes.cast() to cast a float pointer to int pointer, or you can use struct to pack the float value in to a string and unpack it to int. Here is an example:

I use ctypes.cast() to view float value memory as int, and use struct to convert it back:

In [5]: from ctypes import *
In [6]: f = c_float(12.3)
In [8]: tmp = cast(pointer(f), POINTER(c_int)) 
In [9]: tmp.contents
Out[9]: c_long(1095027917)
In [10]: tmp.contents.value
Out[10]: 1095027917
In [11]: hex(tmp.contents.value)
Out[11]: '0x4144cccd'

In [13]: import struct
In [16]: struct.unpack("f",struct.pack("I", 1095027917))
Out[16]: (12.300000190734863,)