Date: Sat, 19 Oct 1996 00:25:53 -0700 (PDT) From: Dmitry Kohmanyuk <dk@dog.farm.org> To: Andrew.Tridgell@anu.edu.au Cc: freebsd-hackers@freebsd.org Subject: Re: fix for symlinks in /tmp (fwd) FYI Message-ID: <199610190725.AAA18425@dog.farm.org>
next in thread | raw e-mail | index | archive | help
In article <96Oct19.133056+1000est.65234-172+1149@arvidsjaur.anu.edu.au> you wrote: > sure, but does it mandate that the rules for when a user can follow a > symlink? That doesn't sound like a POSIX thing to me, more of a > "tradition" thing. I could easily be wrong :-) hmm, I wonder how `symlinks work all the time but' idea is POSIX-compatible ;-) > > The historical BSD behavior is group inheritance, actually, totally > > unrelated to the behaviour needed for the bug (I think). > Hmmm, I thought group inheritance was controlled by the setgid bit on > directories? in 4.4BSD, it's not - it's always this way. It's SYSV-ish idea (which Linux has adopted) to use setgid dirs for `BSD directory semantics. (Although on Linux you can make all your directories this way by using some mount flags for ext2fs (I always did that on all Linux hosts I administered.) > Does the t bit really affect group inheritance in BSD? no, the t bit only means `it if isn't yours, you can't delete it', both in SYSV and 4.4BSD. -- "C makes it easy to shoot yourself in the foot, C++ makes it harder, but when you do, it blows away your whole leg" -- Bjarne Stroustrup
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199610190725.AAA18425>