Bug report

The bounds check for the start_points argument is wrong if the x or y coordinates are not in ascending order:

import matplotlib.pylab as plt
import numpy as np

u = np.ones((2, 2))
v = np.zeros((2, 2))
x = np.array([0, 10])

y = np.array([10, 0])  # (a)
#y = np.array([0, 10])  # (b)

plt.streamplot(x, y, u, v, start_points=[[5, 5]])  # (c)
#plt.streamplot(x, y, u, v)  # (d)

plt.show()

The code results in

ValueError: Starting point (5.0, 5.0) outside of data boundaries

I would have expected that a plot with one streamline is created. The error goes away either if I replace (a) with (b) or if I replace (c) with (d). Thus if start_points are not given, plotting a descending range works totally fine.

Matplotlib version

• Operating system: macOS 10.15.7
• Matplotlib version: 3.3.2
• Matplotlib backend (print(matplotlib.get_backend())): MacOSX
• Python version: 3.8.6
• Other libraries: numpy: 1.19.2

Matplotlib has been installed via pip.