aeb@turing.UUCP (05/29/84)
In contradiction to the manual, the system call setreuid(2) can change
the real uid of a non-superuser process - in fact
setreuid(geteuid(), -1);
works (BSD4.2). Is this desirable?
--
Andries Brouwer -- CWI, Amsterdam -- {philabs,decvax}!mcvax!aebchris@umcp-cs.UUCP (05/31/84)
I would say so; that makes it compatible with setuid under 4.1...
(Specifically, the manual *should* say something like
Setreuid is permitted if and only if
- the real or effective uid is 0, or
- the uid being set is the same as the current real
or effective uid.
Or isn't that what you meant?)
--
In-Real-Life: Chris Torek, Univ of MD Comp Sci (301) 454-7690
UUCP: {seismo,allegra,brl-bmd}!umcp-cs!chris
CSNet: chris@umcp-cs ARPA: chris@maryland