jitclass doesn't use correct type for init arg

[matham]

Reporting a bug

• I have tried using the latest released version of Numba (most recent is
  visible in the release notes
  (https://numba.readthedocs.io/en/stable/release-notes-overview.html).
• I have included a self contained code sample to reproduce the problem.
  i.e. it's possible to run as 'python bug.py'.

The issue is easily demonstrated with this example code:

from numba.experimental import jitclass
from numba import uint64
import numpy as np

spec = [
    ('value', uint64),
]

@jitclass(spec)
class Data(object):
    def __init__(self, value: int):
        self.value = value

If I instantiate the data objects in this order the code works fine:

data = Data(np.iinfo(np.uint64).max)
data = Data(np.iinfo(np.uint32).max)

But, if I do this:

mybag = Data(np.iinfo(np.uint32).max)
mybag = Data(np.iinfo(np.uint64).max)

I get this error:

Traceback (most recent call last):
  File "D:\code\cellcount\cellcount\play3.py", line 15, in <module>
    mybag = Data(np.iinfo(np.uint64).max)
  File "D:\code\cellcount\lib\site-packages\numba\experimental\jitclass\base.py", line 124, in __call__
    return cls._ctor(*bind.args[1:], **bind.kwargs)
OverflowError: int too big to convert

[gmarkall]

Thanks for the report - I can confirm that I can reproduce the issue locally.