urlichs@smurf.sub.org (Matthias Urlichs) (11/07/90)
[ Redirected to news.software.b ]
In news.sysadmin, article <1990Nov6.204258.1833@uokmax.ecn.uoknor.edu>,
rmtodd@uokmax.ecn.uoknor.edu (Richard Michael Todd) writes:
<
< >3) When I first built Cnews for A/UX, I'm sure that I didn't need setnewsids
< > (the setuid root file in /usr/lib/news). Then I fixed some things, changed
< > a few others, and now I can't get Cnews to work without that file. Have I
< > done something obviously stupid here?
< Afraid so :-). A/UX needs setnewsids.
Why? setnewsids is, according to notebook/ctlfiles, only needed when
setuid(geteuid()) doesn't work. On A/UX 2.0, it does work.
So whatever the problem is, that's not it.
--
Matthias Urlichs -- urlichs@smurf.sub.org -- urlichs@smurf.ira.uka.de /(o\
Humboldtstrasse 7 - 7500 Karlsruhe 1 - FRG -- +49+721+621127(0700-2330) \o)/henry@zoo.toronto.edu (Henry Spencer) (11/13/90)
In article <qbv6g2.3wa@smurf.sub.org> urlichs@smurf.sub.org (Matthias Urlichs) writes: >< Afraid so :-). A/UX needs setnewsids. > >Why? setnewsids is, according to notebook/ctlfiles, only needed when >setuid(geteuid()) doesn't work. On A/UX 2.0, it does work. For some value of "work". The problem is that POSIX, and some systems that are moving towards conformance with it, have a different idea of how setuid(geteuid()) should work than the one that used to be widespread. News wants to see that call set both effective and real uids. The new and improved definition doesn't do that. This stuff is slated for revision. -- "I don't *want* to be normal!" | Henry Spencer at U of Toronto Zoology "Not to worry." | henry@zoo.toronto.edu utzoo!henry
urlichs@smurf.sub.org (Matthias Urlichs) (11/16/90)
In news.software.b, article <1990Nov13.032443.9127@zoo.toronto.edu>, henry@zoo.toronto.edu (Henry Spencer) writes: < In article <qbv6g2.3wa@smurf.sub.org> urlichs@smurf.sub.org (Matthias Urlichs) writes: < > < >setnewsids is, according to notebook/ctlfiles, only needed when < >setuid(geteuid()) doesn't work. On A/UX 2.0, it does work. < < For some value of "work". The problem is that POSIX, and some systems < that are moving towards conformance with it, have a different idea of < how setuid(geteuid()) should work than the one that used to be widespread. A/UX, on the other hand, can cope with all of these ideas, depending on which compatibility library you link with (if any). The default, and the BSD compatibility mode, use the old semantics of setuid(geteuid); SVID and POSIX mode use the new one. Guess how I built my Cnews? -- Matthias Urlichs -- urlichs@smurf.sub.org -- urlichs@smurf.ira.uka.de /(o\ Humboldtstrasse 7 - 7500 Karlsruhe 1 - FRG -- +49+721+621127(0700-2330) \o)/