Memory leak when calling copy or deepcopy on any GLM object

[tokotchd]

Hello, I read through #40
and
c10a030

When I run the following code

import tracemalloc
from glm import dvec3
import copy

if __name__ == "__main__":
    test_dvec = dvec3(1, 2, 3)
    tracemalloc.start()
    snapshot1 = tracemalloc.take_snapshot()
    while True:
        copy_test = copy.deepcopy(test_dvec)
        del copy_test
        snapshot2 = tracemalloc.take_snapshot()
        top_stats = snapshot2.compare_to(snapshot1, 'filename')
        print(top_stats[:1], end='\r')

RAM usage increases without bound. If you run an equivalent test against numpy (or other number libraries) the memory usage remains stable:

import tracemalloc
import numpy as np
import copy

if __name__ == "__main__":
    test_np = np.array([1, 2, 3])
    tracemalloc.start()
    snapshot1 = tracemalloc.take_snapshot()
    while True:
        copy_test = copy.deepcopy(test_np)
        del copy_test
        snapshot2 = tracemalloc.take_snapshot()
        top_stats = snapshot2.compare_to(snapshot1, 'filename')
        print(top_stats[:1], end='\r')

[tokotchd]

Update: many more functions are affected including copy, deepcopy, to_tuple, etc.
In the majority of cases, there are simply no calls to dereference copies via Py_DECREF