[Bug]: Setting norm with existing colorbar fails with 3.6.3 · matplotlib/matplotlib challenge

Problem statement

Bug summary

Setting the norm to a LogNorm after the colorbar has been created (e.g. in interactive code) fails with an Invalid vmin value in matplotlib 3.6.3.

The same code worked in previous matplotlib versions.

Not that vmin and vmax are explicitly set to values valid for LogNorm and no negative values (or values == 0) exist in the input data.

Code for reproduction

python

import matplotlib.pyplot as plt
from matplotlib.colors import LogNorm
import numpy as np

# create some random data to fill a 2d plot
rng = np.random.default_rng(0)
img = rng.uniform(1, 5, (25, 25))

# plot it
fig, ax = plt.subplots(layout="constrained")
plot = ax.pcolormesh(img)
cbar = fig.colorbar(plot, ax=ax)

vmin = 1
vmax = 5

plt.ion()
fig.show()
plt.pause(0.5)

plot.norm = LogNorm(vmin, vmax)
plot.autoscale()
plt.pause(0.5)

Actual outcome

Traceback (most recent call last):
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/backends/backend_qt.py", line 454, in _draw_idle
    self.draw()
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/backends/backend_agg.py", line 405, in draw
    self.figure.draw(self.renderer)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/artist.py", line 74, in draw_wrapper
    result = draw(artist, renderer, *args, **kwargs)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/artist.py", line 51, in draw_wrapper
    return draw(artist, renderer)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/figure.py", line 3082, in draw
    mimage._draw_list_compositing_images(
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/image.py", line 131, in _draw_list_compositing_images
    a.draw(renderer)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/artist.py", line 51, in draw_wrapper
    return draw(artist, renderer)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/axes/_base.py", line 3100, in draw
    mimage._draw_list_compositing_images(
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/image.py", line 131, in _draw_list_compositing_images
    a.draw(renderer)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/artist.py", line 51, in draw_wrapper
    return draw(artist, renderer)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/collections.py", line 2148, in draw
    self.update_scalarmappable()
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/collections.py", line 891, in update_scalarmappable
    self._mapped_colors = self.to_rgba(self._A, self._alpha)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/cm.py", line 511, in to_rgba
    x = self.norm(x)
  File "/home/mnoethe/.local/conda/envs/cta-dev/lib/python3.9/site-packages/matplotlib/colors.py", line 1694, in __call__
    raise ValueError("Invalid vmin or vmax")
ValueError: Invalid vmin or vmax

Expected outcome

Works, colorbar and mappable are updated with new norm.

Additional information

_No response_

Operating system

Linux

Matplotlib Version

3.6.3 (works with 3.6.2)

Matplotlib Backend

Multpiple backends tested, same error in all (Qt5Agg, TkAgg, agg, ...)

Python version

3.9.15

Jupyter version

not in jupyter

Installation

conda

The environment starts at commit 66f7956984cb (matplotlib/matplotlib 3.6), dependencies installed and tests runnable from the first minute. You are graded by hidden tests taken from the fix that was actually merged upstream.