[OpenAFS] Very odd pts database issue

[Russ Allbery]

Jeffrey Hutzelman <jhutz@cmu.edu> writes:

> That ID is reserved, to identify objects created by unauthenticated
> users.

Aha!  I have never seen that before.  Might be worth some sort of error
message or explanatory report in pts, since empty pts exam and pts mem
returns is rather disconcerting.

> You'd probably do well to skip 32767 and 32768 as well, for reasons not
> related to AFS.

That ship sailed a long time ago here; we just don't support platforms
that used signed 16-bit UIDs or allocate their nobody users or the like
around that boundary.  I should have recognized the number range, though.

-- 
Russ Allbery (rra@stanford.edu)             <http://www.eyrie.org/~eagle/>