Replay may still do some light disk I/O? #796

When running replay, the result says

This run was a replay from a previous run. No I/O done!

Is that actually true though? I see >700kBps on my USB NTFS drive, with the PC basically idle while rmlint replay is running. Nothing else uses NTFS on this system. The command was run out of an ext4 folder and reading and writing to that folder.

The message is hardcoded for all replays. Does replay do disk I/O?

Maybe a fluke. Re-ran with iotop and didn't see anything - but the results completed faster (45s vs 180s) suggesting maybe some caching.