Unexpected drop_log behaviour when subsetting epochs

Description of the problem

Bad epoch from irrelevant condition isn't marked as "IGNORED" in drop_log.

Expected results

"IGNORED" is added to the irrelevant epochs even when they're bad.

Actual results

For example: I have only two conditions, either 'stim' or 'blank'.
For epochs['stim'].drop_log the output is:
(('IGNORED',), (), ('EEG044',), (), (), ...
While epochs['blank'].drop_log returns:
((), ('IGNORED',), ('EEG044',), ('IGNORED',), ('IGNORED',), ...

Note how the third epoch isn't ignored in either condition.

Additional information

Platform Windows-11-10.0.22631-SP0
Python 3.12.7 (tags/v3.12.7:0b05ead, Oct 1 2024, 03:06:41) [MSC v.1941 64 bit (AMD64)]
Executable c:\Users\gennadiyb\Documents\erp.venv\Scripts\python.exe
CPU Intel64 Family 6 Model 158 Stepping 10, GenuineIntel (12 cores)
Memory 31.9 GB

Core
├☑ mne 1.8.0 (latest release)
├☑ numpy 2.1.2 (OpenBLAS 0.3.27 with 12 threads)
├☑ scipy 1.14.1
└☑ matplotlib 3.9.2 (backend=module://matplotlib_inline.backend_inline)

Numerical (optional)
├☑ pandas 2.2.3
└☐ unavailable sklearn, numba, nibabel, nilearn, dipy, openmeeg, cupy, h5io, h5py

Visualization (optional)
├☑ qtpy 2.4.1 (PyQt6=6.7.1)
├☑ pyqtgraph 0.13.7
├☑ mne-qt-browser 0.6.3
├☑ ipywidgets 8.1.5
└☐ unavailable pyvista, pyvistaqt, vtk, ipympl, trame_client, trame_server, trame_vtk, trame_vuetify

Ecosystem (optional)
└☐ unavailable mne-bids, mne-nirs, mne-features, mne-connectivity, mne-icalabel, mne-bids-pipeline, neo, eeglabio, edfio, mffpy, pybv

Start by reproducing the behavior with epochs['stim'].drop_log and epochs['blank'].drop_log, then trace the epoch-subsetting and drop_log handling paths. Add a regression test covering bad epochs from irrelevant conditions, and consider the issue done when irrelevant epochs are marked "IGNORED" in both subsets.