plotting cov from mne.make_ad_hoc_cov leads to error

I want to create an identity covariance matrix. Using the code (make_ad_hoc_cov) below seems to work, however plotting leads to:

IndexError: too many indices for array: array is 1-dimensional, but 2 were indexed

Going the non-recommended way of instantiating mne.Covariance directly works, however.

Am I on the wrong track to create an identity matrix this way, or is this a potential issue with the code in mne?

import mne
import numpy as np

n = 50
sfreq = 100
info = mne.create_info([f"ch{i}" for i in range(n)], sfreq, ch_types='eeg', verbose=None)

# recommended way?
cov = mne.cov.make_ad_hoc_cov(info, std={"eeg": 1})
cov.plot(info)  # error

# not recommended
cov2 = mne.Covariance(data=np.eye(n), names=info.ch_names, bads=info["bads"], projs=[], nfree=1)
cov2.plot(info)  # works

Start by reproducing the example with mne.cov.make_ad_hoc_cov(info, std={"eeg": 1}) and cov.plot(info), then compare it with the directly constructed mne.Covariance example that plots successfully. Trace the make_ad_hoc_cov and Covariance.plot entry points; done means the recommended covariance object plots without the IndexError and the existing behavior remains covered by a regression test.