Report which custom reference was applied

Currently, after calling set_eeg_reference() with a custom channel name, or invoking add_reference_channels() (which calls the former after adding all-zeros channels), to my knowledge it's impossible to retrieve which reference channels are being used in the current referencing scheme. The only thing that gets set in info is 'custom_ref_applied'.

import os.path as op
import mne

data_path = mne.datasets.testing.data_path()
examples_dir = op.join(data_path, 'Brainvision')
vhdr_file = op.join(examples_dir, 'Analyzer_nV_Export.vhdr')

raw = mne.io.read_raw(vhdr_file, preload=True)
raw.set_eeg_reference(['F8'])

# do fancy stuff here
# ...
# Wait … WHICH reference did I apply again? Guess I'll have to find the channel
# with all-zero data :(

Isn't there a way to keep track of which references were set?

Start by tracing the set_eeg_reference() and add_reference_channels() entry points and how they update info, then review the issue discussion for the expected metadata behavior. Done means the custom reference channel names remain retrievable after applying either operation and subsequent processing.