Greetings!

First, thanks to the NumPy community for your efforts.

Second, In NumPy 1.16.5 I observe the following:

>>> import numpy as np
>>> np.clip(2, 3, 0)
3

In NumPy 1.17.4, I get a different result:

>>> import numpy as np
>>> np.clip(2, 3, 0)
0

I see that the difference has to do with which constraint gets precedence. Was this an intentional change in precedence then?

Reproducing code example:

import numpy as np
assert np.clip(2, 3, 0), 3

Numpy/Python version information:

>>> import sys, numpy; print(numpy.__version__, sys.version)
1.17.4 3.7.4 (default, Dec  3 2019, 11:37:16) 
[GCC 5.4.0 20160609]