[ishaaaan17]

summary of changes

fixes #9953.
this builds on and finishes the abandoned PR #10970 by @TzionaN, rebased onto the current main and adapted to mne/_fiff/meas_info.py.

before this, set_meas_date raised a ValueError if you gave it a timezone-aware datetime object with a non-UTC timezone, while setting info['utc_offset'] directly was deprecated.

this change does these things:

1. set_meas_date now accepts timezone-aware non-UTC datetime objects, converts them to UTC for info['meas_date'], and fills info['utc_offset'] automatically with the formatted offset (for example +02:00 or -05:00).
2. sets info['utc_offset'] = None when meas_date is in UTC or None.
3. adds parsing of the offset in ISO8601 strings in _handle_meas_date.
4. adds unit tests in mne/_fiff/tests/test_meas_info.py (including a FIFF I/O save/reload roundtrip and anonymization) and in mne/tests/test_annotations.py.
5. adds a changelog entry in doc/changes/dev/9953.bugfix.rst.

Demonstration

from datetime import datetime, timezone, timedelta
import mne

info = mne.create_info(ch_names=["EEG 001"], sfreq=1000.0, ch_types="eeg")
tz_paris = timezone(timedelta(hours=2))
dt = datetime(2026, 10, 20, 14, 30, tzinfo=tz_paris)

# main: ValueError: Date must be a timezone-aware datetime object: ...
# fix: converts to UTC for meas_date and sets utc_offset = "+02:00"
info.set_meas_date(dt)
print("meas_date:", info["meas_date"])   # 2026-10-20 12:30:00+00:00
print("utc_offset:", info["utc_offset"]) # +02:00

ai disclosure

i used an ai assistant to trace how utc_offset is formatted in FIFF and how Info.set_meas_date is structured across mne/_fiff/meas_info.py and mne/annotations.py. i checked the offset calculation logic, wrote the round-trip and anonymization tests, formatted the code with ruff, and ran the pytest suite locally before pushing.

[larsoner]

Pushed a minor commit for a bug this made me notice that we had in how we formatted the offset for EGI data. Marking for merge-when-green, thanks @ishaaaan17 !