Thanks for confirming! It's probably not directly something we can fix, but I'll check with the ORM team if we could at least get a more explicit failure.