There is one change in behaviour I noticed when upgrading to Python 3.8, where matplotlib 3.1.1 used to work in Python 3.7, but not in Python 3.8 any more.

>>> import numpy as np
>>> import matplotlib.pyplot as plt
>>> plt.xlim(np.array([3]), 4)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/deil/opt/anaconda3/envs/py38/lib/python3.8/site-packages/matplotlib/pyplot.py", line 1434, in xlim
    ret = ax.set_xlim(*args, **kwargs)
  File "/Users/deil/opt/anaconda3/envs/py38/lib/python3.8/site-packages/matplotlib/axes/_base.py", line 3270, in set_xlim
    left, right = sorted([left, right], reverse=reverse)
TypeError: only integer scalar arrays can be converted to a scalar index

I think the issue is that reverse = np.array([False]) and sorted refuses to process that, raising the TypeError. I'm not sure if this is a change in behaviour in CPython, or if I'm having slightly different Numpy & matplotlib versions in my Python 3.7 and 3.8 environments.

In our scripts we have code that computes the left and right for set_xlim, and for complex reasons it gives a 1-element ndim=1 Numpy array, not a Numpy scalar.

Now the docstring says that only scalar is supported:
https://matplotlib.org/3.1.1/api/_as_gen/matplotlib.axes.Axes.set_xlim.html

Is the old behaviour to work OK with single-entry ndim=1 arrays something you want to support via some code change in matplotlib?
Or is this a usage bug that only accidentally worked earlier, and we fix our code?