>> Look in curproc->p_session->s_login
>
>Hmm!  Of course, the obvious next question is "what if the process
>doesn't belong to a session?" (or, if that's not possible, what if it
>belongs to the same session as init and inetd and such?).

then it would be root, if such a value was set.  we really shouldn't
be getting that picky about which sessions get used and which ones
don't.  it should all be the same.

>Nevertheless, at the very least, we now have two potential formats: one
>for this string (p->p_session->s_login) and one for the process's UID
>(p->p_ucred->cr_uid or perhaps p->p_cred->pc_ruid - or perhaps two
>formats, one for each).

hmm...i'd suggest using pc_ruid.  of course...it shouldn't actually
leave a core file behind if the two values are different, eh?

>What happens for the login user string if no session login name has
>been set would have to be specified.

fall back to pc_ruid?

-- 
|-----< "CODE WARRIOR" >-----|
codewarrior@daemon.org             * "ah!  i see you have the internet
twofsonet@graffiti.com (Andrew Brown)                that goes *ping*!"
andrew@crossbar.com       * "information is power -- share the wealth."