afs_strerror()?

[John Hawkinson]

I'm sort of wondering if we should have one.
Or perhaps an rx_strerror().

Because I haven't gone and followed up on Assar's
suggestions:

| It would seem to me that you need to fix all the arlalib_* functions
| to take the cell as an argument.

(I'll try to do it today), pts on the athena cell doesn't work:

> ./pts mem jhawk
mem: pts: pts_name_to_id(..) returned 19270408

Of course, this is 

include/rxkad.h:#define RXKADUNKNOWNKEY         (19270408L)

But the pts code just does:

        if(res != 0) {
            if(res == PRPERM) {
                if(id == ANONYMOUSID) /* this sucks */
                    errx(1, "pts: User or group doesn't exist so couldn't"
                         "look up id for %s (id: %d)", name, id);
                else
                    errx(1, "pts: Permission denied ; unable to get "
                         "membership of %s (id: %d)", name, id);
            }
            else if(res == PRNOENT)
                errx(1, "pts: User or group doesn't exist so couldn't look up"
                     "id for %s (id: %d)", name, id);
            else
                errx(1, "pts: PR_ListElements(..) returned %d", res);
        }

which is clearly insufficient since all sorts of errors
could pop up.

An rx_strerror() seems like it'd be the right thing here.
Concurrance? Disagreement?

Thanks.

--jhawk