Example log: https://github.com/astropy/astropy/actions/runs/7392979516/job/20112234044

Things in RC:

• pandas 2.2.0rc0
• scipy 1.12.0rc1
• pytest 8.0.0rc1

Since we test against scipy-dev and pytest-dev (in different jobs though), likely culprit is pandas, but we might have also overlooked something that snuck in any of the dev during the holidays.

1. io/fits/tests/test_util.py:27 with Failed: DID NOT WARN.
2. table/tests/test_mixin.py:170 with AstropyUserWarning: Earth Location "TOPOCENTER" for Time Column "atcb" is incompatible with scale "TCB". (2 similar failures)
3. timeseries/io/tests/test_kepler.py:107 with UnitsWarning: 'BJD - 2457000, days' did not parse as fits unit
4. timeseries/tests/test_sampled.py:399 with UnitsWarning: 'BJD - 2457000, days' did not parse as fits unit
5. utils/tests/test_data.py:1101 with CacheMissingWarning: Remote data cache could not be accessed due to OSError

Some of these also look suspicious. I wonder if there was a remote network failure and these are transient. I restarted the job and all the failures remain. Whoever wants to investigate will have to install pre-release(s) of suspected package(s) and see if you can reproduce the errors.