mne-tools / mne-bids-pipeline

Automatically process entire electrophysiological datasets using MNE-Python.
https://mne.tools/mne-bids-pipeline/
BSD 3-Clause "New" or "Revised" License
134 stars 65 forks source link

MAINT: Test result caching #836

Closed larsoner closed 6 months ago

larsoner commented 6 months ago

Before merging …

Closes #814

Let's see if we can replicate the issue! EDIT: Spoiler alert, we cannot.

larsoner commented 6 months ago

@hoechenberger this appears headed for green. I was unable to replicate the caching issue on macOS, both datasets tested here appear to process in < 15s the second time. Ready for review/merge from my end. Happy to adjust CI requirements if you're happy (I folded the Style check into the doc consistency one since the only step we really needed there now was codespell, and knew we'd be adjusting requirements here anyway!)

hoechenberger commented 6 months ago

Thanks! I also realized I cannot reliably reproduce on macOS. Sometimes the problem occurs and sometimes it doesn't. I usually work inside of a OneDrive folder, maybe this has implications...

larsoner commented 6 months ago

I usually work inside of a OneDrive folder, maybe this has implications...

It seems like it could for mtime but for hash it would be weird. It really seems like the file hash shouldn't change!

hoechenberger commented 6 months ago

Let's just go ahead here and I'll report back sometime in the future if I run into this issue again :)