@User OpenRAM uses LEF views for the bitcells which remove the DRC violations that are allowed in memory cells. So you don't have 2M errors.

that appears to have stopped working for me

You are using it in the OpenLane flow. I can't really say what they've changed there.

I only run DRC checks in OpenLane and the precheck

Yeah, so that is not something I control..

I understand

yep, there seems to be some debate in https://github.com/RTimothyEdwards/open_pdks/pull/220#pullrequestreview-891911509 currently if the fix should go in open_pdks or in the precheck themselves

but the thing that is actually blocking me is the klayout FEOL errors in the precheck

I think I agree with @User that if they decide to explicitly skip those checks, that should go in the precheck rather than open_pdks.

which I think is a real issue and you're aware of it and are working on a fix - right?

Yes, those weren't previously checked. It is fixed but I haven't pushed because the DRC/LVS is not passing due to other versioning issues

good luck