Colorbar for imshow messes interactive cursor with masked data

[Descanonge]

Bug report

Adding a colorbar when the mappable is the result of imshow for a masked array changes the behavior of the interactive z-data cursor and triggers a UserWarning.

Code for reproduction

import numpy as np
import matplotlib.pyplot as plt

mask = np.zeros((4, 4), bool)
mask[:2] = True
A = np.ma.array(np.random.rand(4, 4), mask=mask)

fig, ax = plt.subplots(1, 1)
im = ax.imshow(A)
fig.colorbar(im)

Actual outcome

~/.anaconda3/lib/python3.7/site-packages/matplotlib/ticker.py:632: UserWarning: Warning: converting a masked element to nan.
  return '%-12g' % value

Expected outcome
Before adding the colorbar, the zdata cursor shows a blank ([]).

Matplotlib version

• Operating system: Linux
• Matplotlib version: 3.1.1
• Matplotlib backend (print(matplotlib.get_backend())): Qt5Agg
• Python version: 3.7.3
• IPython version: 7.2.0
• Other libraries:
  Matplotlib installed from pip.

[jklymak]

Any idea if this is a change of behaviour from previous versions? I know that the formatter changes to use the colorbar's formatter if its available, but not sure when that was introduced.

[Descanonge]

I don't remember this warning ever coming up until recently.
After a bit of rollback, it seems this issue appeared with 3.1.0.

[ImportanceOfBeingErnest]

Bisects to #12459