Phew! Found the illusive bug in my NGINX robot access project.
It turns out the robotstxt dependency was intended for use by web crawlers rather than web servers. For details, see "An unexpected bug" in the post (https://underlap.org/nginx-robot-access).
Add comment