Numpy Issue with Bigger bit sizes

[DaveKirkland]

Getting strange results when using larger bit sizes in numpy calls

# This works
w = Fxp([1, 1, 1, 1], dtype='fxp-s29/0')
y=np.cumsum(w)
y.info()

# This doesn't
w = Fxp([1, 1, 1, 1], dtype='fxp-s32/0')
y=np.cumsum(w)
y.info()

The first call returns [1 2 3 4], while the 2nd returns [-1 -1 -1 -1]

Version info
OS - Windows 10
Python - 3.8.5
Numpy - 1.21.2
fxpmath - 0.4.8

[francof2a]

solved in v0.4.9 for linux and MAC OS.
Support for windows still under review