Sorry, that last PR ended up way in the weeds. I convinced myself the PR was an improvement, and it was really just uncovering an existing inconsistency that we should tackle separately. https://github.com/osquery/osquery/issues/7945