Note that this issue doesn`t occur when using matplotlib

import numpy as np
import matplotlib.pyplot as plt

fig = plt.figure()

ax = plt.subplot()
ax.set_xlim(-2, 2)
ax.set_ylim(-2, 2)
ax.set_title('line plot A->B vs B->A')

ax.set_xlabel('x axis')
ax.set_ylabel('y axis')

ax.scatter(-1, 1, color="b")
ax.text(-1, 1, "A")
ax.scatter(1, -1, color="b")
ax.text(1, -1, "B")

#x_set = [-1, 1]
#y_set = [1, -1]

x_set = [1, -1]
y_set = [-1, 1]

ax.plot(x_set, y_set, color="r")
pplt.show()

This is part of the "standardized arguments" feature (see this link)... but unfortunately I did not document it! I'll add this to the docs in the next version.

I find that most of the time, when I have "reversed" dependent variable coordinates, I want that axis to be reversed too. So I think I'd prefer to keep this default behavior. In your example, you can use ax.format(xreverse=False) to reverse the axis back. Alternatively (this is currently not part of the public API) ax.plot(x_set, y_set, autoreverse=False) should also work. Might make that part of the public API in the future.

Ok, I understand your point. Thanks for the reply

No problem. FYI the master branch now includes documentation on auto-reversing in the "1D plotting" and "2D plotting" sections, and auto-reversing is now automatically disabled whenever the axis limits were previously explicitly fixed with e.g. ax.format(xlim=xlim) or ax.set_xlim(xlim) (f34c12c).