Fix Issue #326: Adapt to PyLongObject changes in Python 3.12+

[Komorebi-Z]

In Python 3.12 and newer, the PyLongObject struct was changed, making Py_SIZE() unreliable for getting the length of the digit array. This issue caused longObjToMPZ() to incorrectly determine the length of long_obj during the initialization function Element_elem() in charm/core/math/pairing/pairingmodule.c, leading to incorrect conversion behavior.

The following changes have been implemented:

1. Introduced PyLongObj_Size macro to calculate the size of PyLongObject struct.
2. Renamed PythonLongVal to PyLongObj_Val to improve naming consistency.

This probably fixes issue #326, but further testing on different Python versions is recommended for caution.

Note: Further adjustments may be required for Python 3.14+ due to potential shifts to new C APIs, as detailed in PEP 757.

[BatchClayderman]

Hi, guys. Thank you so much for fixing those issues in ZR.
As I am not quite familiar with PairingGroup operations, I am also interested in whether the following operations of G1 elements are outputting correct results in Python 3.12. Thanks a lot.

Python 3.12.3 (main, Feb  4 2025, 14:48:35) [GCC 13.3.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from charm.toolbox.pairinggroup import PairingGroup, G1, G2, GT, ZR, pair, pc_element as Element
>>> group = PairingGroup("MNT159")
>>> a = group.random(G1)
>>> a
[301187472002304746619209247963206559947998705076, 121731044715532841670820951857195078999418162]
>>> group.init(G1, 1)
O
>>> group.init(G1, 2)
O
>>> group.init(G1, 3)
O
>>> a * group.init(G1, 1)
[301187472002304746619209247963206559947998705076, 121731044715532841670820951857195078999418162]
>>> a * group.init(G1, 2)
[301187472002304746619209247963206559947998705076, 121731044715532841670820951857195078999418162]
>>> a * group.init(G1, 3)
[301187472002304746619209247963206559947998705076, 121731044715532841670820951857195078999418162]
>>>

[Komorebi-Z]

@BatchClayderman Don't worry, this is normal and depends on the implementation of the set_mpz method in different groups. Taking PBC as an example, group.init(G1, 2) eventually calls zero_set_mpz located here, which is designed to ignore the input value.

[jakinyele]

This is a good catch. Thanks for the PR @Komorebi-Z.