SEP 21: Reject relative URLS

I don’t use relative URLS on this site for reasons of robustness: in the event that a document moved (or a section of a document containing a relative URL was copied into another document) that relative URL would now point somewhere other than where it was intended.

While implementing ✅ SEP 16: Extract backlinks from plaintext I did happen to discover one dormant relative URL in an old document. It pointed nowhere (404) because it was a correctly formed absolute URL except it was missing its leading ‘/’. I fixed it up, glad that my implementation of ✅ SEP 16: Extract backlinks from plaintext had thrown a KeyError, but realised that no all accidental (relative) URLs would be caught by this system in the future.

Given my preference for never using them it makes sense to have the site build throw an error if it encounters one anywhere.

Another example where an accidental relative URL could appear is in any case where I forgot to precede a URN reference ✅ SEP 12: Reference documents by their URN/UUID with a ‘!’.