Bug summary

I'm plotting some log-scale data with ax.errorbar. If I plot first and then call ax.set_[x/y]scale, autolimits work as expected. However, switching the order (first setting the scale, then plotting the errorbars) leads to incorrect autolimits.

Code for reproduction

import numpy as np
from matplotlib import pyplot as plt

x = 10 ** np.array([18, 18.1, 18.2, 18.3])
y = np.array([100, 80, 60, 30])
yerr = np.ones_like(y) * 10

fig, ax = plt.subplots()
ax.set_xscale("log")
ax.set_yscale("log")
ax.errorbar(x, y, yerr=yerr)

fig.savefig("test.png")

Actual outcome

Both x and y axes have lower limits at 1. Upper limits seem correct.

Expected outcome

If I call set_*scale after plotting, everything works as expected:

The same plot should be produced by the MRE script above.

Additional information

No response

Operating system

Ubuntu

Matplotlib Version

3.10.8, also occurred on 3.10.5

Matplotlib Backend

agg

Python version

3.13.7

Jupyter version

No response

Installation

pip