incorrect default limits 'log' scale depending on order of scatter #11198

import numpy as np
import matplotlib.pyplot as plt

# smallest span
x1 = np.array([1, 10])
y1 = np.array([10, 1])

# largest span
x2 = np.array([0.1, 100])
y2 = np.array([100, 0.1])

fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(2,2, figsize=(8,8))

ax1.scatter(x1, y1)
ax1.scatter(x2, y2)
ax1.set_title('linear; smallest span first', color='g')

ax2.scatter(x1, y1)
ax2.scatter(x2, y2)
ax2.set_xscale('log')
ax2.set_yscale('log')
ax2.set_title('log; smallest span first', color='r')

ax3.scatter(x2, y2)
ax3.scatter(x1, y1)
ax3.set_title('linear; largest span first', color='g')

ax4.scatter(x2, y2)
ax4.scatter(x1, y1)
ax4.set_xscale('log')
ax4.set_yscale('log')
ax4.set_title('log; largest span first', color='g')