50.2 argmax, argmin, min, max weird behaviour int8 and uint8 datatypes

[mz3-andrey]

Am running 50.2 on my openmv's now, but my code started doing weird things.
I looked further into it and noticed min, max, argmin and argmax were the culprit.
Issue #2984 mentions a fix as they weren't working for int16's before, but now only int16s seem to work.

My test:

print(min, max, argmin, argmax) of a given array.
dtypes of float, and uint16 and int16 work exactly as expected.
But for:
[-2, 4, 2, -1, 6, 1, 0, -1], dtype=np.int8 the print result is -2, 116, 0, 5
and for
[5, 7, 1, 4, 8], dtype=np.uint8 the print result is 0, 5, 2, 0

[v923z]

Thanks for reporting the issue! Something is clearly wrong here, I will look into it.

[v923z]

@mz3-andrey #130 fixes the issue.

@jepler Jeff, could you, please, pull this into circuitpython? You can use the https://github.com/v923z/micropython-ulab/tree/0.51.1 tag.

[mz3-andrey]

Great thanks! I appreciate it.